BOWLING ALONE

by Robert D. Putnam

In some cases, like neighborhood lawn raking, the return of the favor is immediate and the calculation straightforward, but in some cases the return is long-term and conjectural, like the benefit of living in the kind of community where people care for neglected children. At this extreme, generalized reciprocity becomes hard to distinguish from altruism and difficult to cast as self-interest. Nevertheless, this is what Tocqueville, insightfully, meant by “self-interest rightly understood.”

  When each of us can relax her guard a little, what economists term “transaction costs”—the costs of the everyday business of life, as well as the costs of commercial transactions—are reduced. This is no doubt why, as economists have recently discovered, trusting communities, other things being equal, have a measurable economic advantage.4 The almost imperceptible background stress of daily “transaction costs”—from worrying about whether you got back the right change from the clerk to double-checking that you locked the car door—may also help explain why students of public health find that life expectancy itself is enhanced in more trustful communities.5 A society that relies on generalized reciprocity is more efficient than a distrustful society, for the same reason that money is more efficient than barter. Honesty and trust lubricate the inevitable frictions of social life.

  “Honesty is the best policy” turns out to be a wise maxim rather than a mawkish platitude, but only if others follow the same principle. Social trust is a valuable community asset if—but only if—it is warranted. You and I will both be better off if we are honest toward one another than if—each fearing betrayal—we decline to cooperate. However, only a seeker of sainthood will be better off being honest in the face of persistent dishonesty. Generalized reciprocity is a community asset, but generalized gullibility is not.6 Trustworthiness, not simply trust, is the key ingredient.7

  In a society of fallible humans, what kind of assurance can each of us have in the good faith of others? A legal system, complete with courts and law enforcement, provides one strong answer. However, if we needed legal advice and a police presence to formulate and enforce the simplest agreement—like whether to rake our respective lawns or share Sunday snack duties—escalating transaction costs would surely preclude much mutually beneficial cooperation. As Diego Gambetta, a student of trust (and of the Mafia), points out, “Societies which rely heavily on the use of force are likely to be less efficient, more costly, and more unpleasant than those where trust is maintained by other means.”8

  Another solution, social science has recently recognized, inheres in the social fabric in which our daily transactions are embedded.9 An effective norm of generalized reciprocity is bolstered by dense networks of social exchange. If two would-be collaborators are members of a tightly knit community, they are likely to encounter one another in the future—or to hear about one another through the grapevine. Thus they have reputations at stake that are almost surely worth more than gains from momentary treachery. In that sense, honesty is encouraged by dense social networks.

  There is an important difference between honesty based on personal experience and honesty based on a general community norm—between trusting Max at the corner store because you’ve known him for years and trusting someone to whom you nodded for the first time at the coffee shop last week. Trust embedded in personal relations that are strong, frequent, and nested in wider networks is sometimes called “thick trust.”10 On the other hand, a thinner trust in “the generalized other,” like your new acquaintance from the coffee shop, also rests implicitly on some background of shared social networks and expectations of reciprocity.11 Thin trust is even more useful than thick trust, because it extends the radius of trust beyond the roster of people whom we can know personally.12 As the social fabric of a community becomes more threadbare, however, its effectiveness in transmitting and sustaining reputations declines, and its power to undergird norms of honesty, generalized reciprocity, and thin trust is enfeebled.

  Referring to what I have labeled “thin trust,” political scientists Wendy Rahn and John Transue observe that “social, or generalized, trust can be viewed as a ‘standing decision’ to give most people—even those whom one does not know from direct experience—the benefit of the doubt.”13 Social trust in this sense is strongly associated with many other forms of civic engagement and social capital. Other things being equal, people who trust their fellow citizens volunteer more often, contribute more to charity, participate more often in politics and community organizations, serve more readily on juries, give blood more frequently, comply more fully with their tax obligations, are more tolerant of minority views, and display many other forms of civic virtue. Moreover, people who are more active in community life are less likely (even in private) to condone cheating on taxes, insurance claims, bank loan forms, and employment applications. Conversely, experimental psychologists have shown that people who believe that others are honest are themselves less likely to lie, cheat, or steal and are more likely to respect the rights of others. In that sense, honesty, civic engagement, and social trust are mutually reinforcing.14

  Figure 37: Declining Perceptions of Honesty and Morality, 1952–1998

  In short, people who trust others are all-round good citizens, and those more engaged in community life are both more trusting and more trustworthy. Conversely, the civically disengaged believe themselves to be surrounded by miscreants and feel less constrained to be honest themselves. The causal arrows among civic involvement, reciprocity, honesty, and social trust are as tangled as well-tossed spaghetti. Only careful, even experimental, research will be able to sort them apart definitively.15 For present purposes, however, we need to recognize that they form a coherent syndrome.

  For all these reasons, an important diagnostic test for trends in social capital in America in recent decades is how reciprocity and social trust have evolved—not merely thick trust in people whom we know intimately, but thin trust in the anonymous other. The central question in this chapter is this: How are the trends in social capital and civic engagement that we have already discovered reflected in trends in honesty and social trust in America?

  OUR SUBJECT HERE is social trust, not trust in government or other social institutions. Trust in other people is logically quite different from trust in institutions and political authorities. One could easily trust one’s neighbor and distrust city hall, or vice versa. Empirically, social and political trust may or may not be correlated, but theoretically, they must be kept distinct. Trust in government may be a cause or a consequence of social trust, but it is not the same thing as social trust.16

  Figure 38: Four Decades of Dwindling Trust: Adults and Teenagers, 1960–1999

  Fortunately, pollsters have been asking Americans standard questions about social trust and honesty for many decades. Unfortunately, the responses contain an irreducible element of ambiguity. Take, for example, the most common survey question: “Generally speaking, would you say that most people can be trusted, or that you can’t be too careful in dealing with people?” This question clearly taps feelings about the trustworthiness of the generalized other—thin trust17—but the meaning of the responses remains murky in one respect. If fewer survey respondents nowadays say, “Most people can be trusted,” that might mean any one of three things: 1) the respondents are accurately reporting that honesty is rarer these days; or 2) other people’s behavior hasn’t really changed, but we have become more paranoid; or 3) neither our ethical demands nor other people’s behavior have actually changed, but we now have more information about their treachery, perhaps because of more lurid media reports.

  Figure 39: Generational Succession Explains Most of the Decline in Social Trust

  It is not easy to sort out what’s going on here, any more than when your kindergartner complains that a playmate acted unfairly. However, the social geography of social trust suggests that survey reports about honesty and trust should be interpreted prima facie as accurate accounts of the respondents’ social experiences. In virtually all societies “have-nots” are
less trusting than “haves,” probably because haves are treated by others with more honesty and respect.18 In America blacks express less social trust than whites, the financially distressed less than the financially comfortable, people in big cities less than small-town dwellers, and people who have been victims of a crime or been through a divorce less than those who haven’t had these experiences.19 It is reasonable to assume that in each case these patterns reflect actual experience rather than different psychic predispositions to distrust. When such people tell pollsters that most people can’t be trusted, they are not hallucinating—they are merely reporting their experience.

  Take, for instance, the case of city size. As we noted in the previous chapter, virtually all forms of altruism—volunteerism, community projects, philanthropy, directions for strangers, aid for the afflicted, and so on—are demonstrably more common in small towns. Crime rates of all sorts are two or three times higher in cities. (Not surprisingly, victims of crime and violence—wherever they live—express reduced social trust, a perfectly intelligible updating of their views about the trustworthiness of others.) Store clerks in small towns are more likely to return overpayment than their urban counterparts. People in small towns are more likely to assist a “wrong number” phone caller than urban dwellers. Cheating on taxes, employment forms, insurance claims, and bank loan applications are three times more likely to be condoned in cities than in small towns. Car dealers in small towns perform far fewer unnecessary repairs than big-city dealerships.20

  In short, the somewhat greater mistrust of the generalized other expressed by residents of big cities is not some peculiar paranoia that arises from urban living, but a realistic account of their actual experience and of social norms in their surroundings. To be sure, weaker informal social control in cities also makes them freer places to live—“City air liberates,” as the medieval proverb had it. Enfeebled thin trust may be a fair price for that freedom. Nevertheless, when urbanites express social distrust, they are accurately reporting something about their social environment.21

  To be sure, social distrust is not purely objective. It also to some extent reflects personal cynicism, paranoia, and even projections of one’s own dishonest inclinations.22 People who feel themselves to be untrustworthy are less trusting of others.23 In fact, social trust can easily generate vicious spirals (or virtuous circles), as my expectations of others’ trustworthiness influences my trustworthiness, which in turn influences others’ behavior. We should begin, however, with the simpler presumption that both those who report that “most people are honest” and those who say that “you can’t be too careful” are sincerely summarizing their own experiences. It is reasonable to suppose, too, that views about something as basic as reciprocity and generalized trust are especially influenced by personal experience and social customs early in life. That, after all, is why we call them the “formative” years.

  Most Americans today believe that we live in a less trustworthy society than our parents did.24 In 1952, as figure 37 shows, Americans were split about fifty-fifty on the issue of whether our society was then as upright morally as it had been in the past. In 1998, however, after nearly four decades of growing cynicism, we believe by a margin of three to one that our society is less honest and moral than it used to be. But perhaps that only proves that nostalgia is in fashion.

  Survey archives allow us to screen out that “golden glow,” at least to some extent, by comparing our feelings today, not with how we imagine an earlier generation might have felt, but rather with what that generation actually said in response to identical questions. The best evidence suggests that social trust rose from the mid-1940s to the mid-1960s, peaking in 1964 just as many other measures of social capital did. Middle-aged Americans in the 1960s were probably living in a more trusting society than the one in which they had grown up.25

  In the mid-1960s, however, this beneficent trend was reversed, initiating a long-term decline in social trust.26 (See figure 38.) Every year fewer and fewer of us aver that “most people can be trusted.” Every year more and more of us caution that “you can’t be too careful in dealing with people.” If generalized reciprocity and honesty are important social lubricants, Americans today are experiencing more friction in our daily lives than our parents and grandparents did a generation ago. As figure 38 makes plain, this decline in social trust has been even steeper among younger Americans than among the rest of us, especially since about 1985.27

  Figure 40: The Changing Observance of Stop Signs

  Most, if not all, of the decline in American social trust since the 1960s is attributable to generational succession.28 Moreover, that generational decline has tended to accelerate in the last decade or two. In the 1970s roughly 80 percent of Americans born in the first third of the century believed that “most people are honest,” and in the late 1990s they continued to hold that optimistic view in almost undiminished degree. (See figure 39.) However, their share in the population had fallen from nearly one in every two adults in 1975 to barely one in every eight adults in 1998. At the same time, in the 1970s roughly 75 percent of those born between 1930 and 1945 believed in the essential honesty of others, and their views also changed little over subsequent decades. Roughly 60 percent of the baby boomers (born 1946–1960) agreed in the 1970s that “most people are honest,” and their views were unchanged in the late 1990s. Finally, at the bottom of the generational hierarchy Americans born after 1960 were not out of adolescence in the mid-1970s, but ever since their cohort began to reach adulthood in the mid-1980s, roughly half of them have denied that “most people are honest.” By 1999 this mistrustful younger generation already constituted nearly one-third of the adult population.

  Figure 41: U.S. Crime Rates, 1960–1997

  Examining the views of the postboomers in finer detail only reinforces the picture of accelerating generational decline. By 1998–99 respondents born in the 1970s—who had been less than five years old when this series of surveys started—constituted a rapidly rising 10 percent of the total population, and only 40 percent of them agreed that “most people are honest.” In short, at century’s end, a generation with a trust quotient of nearly 80 percent was being rapidly replaced by one with a trust quotient of barely half that. The inevitable result is steadily declining social trust, even though each individual cohort is almost as trusting as it ever was.

  On the interpretive assumption that expressions of social trust are, in the first instance, reflections of personal experience, weighted perhaps by early impressions, the social distrust among America’s youth should be seen not as a character flaw, but rather as a mirror held up to social mores of recent decades. Our youth are, in effect, telling us that in their experience most people really aren’t trustworthy. Perhaps thick trust—confidence in personal friends—is as strong as ever, as some Gen X’ers believe. However, thin trust—the tenuous bond between you and your nodding acquaintance from the coffee shop, that crucial emollient for large, complex societies like ours—is becoming rarer.

  Figure 42: Employment in Policing and the Law Soared after 1970

  The evidence of declining generalized trust and reciprocity also shows up in refusal rates for opinion surveys themselves, which have more than doubled since the 1960s. Cooperation rates may have risen slightly from the 1940s to the 1960s—perhaps not coincidentally the same period in which many of the other indicators of social trust and social capital were rising—but response rates were certainly declining by the 1970s. The most exhaustive recent study of survey response rates confirms the trend and adds that social disconnectedness seems to be part of the reason.29 Intriguingly, the rise in refusals in recent years has plagued face-to-face and telephone interviews, but not mail surveys. This pattern suggests that these refusals may be due more to the vague menace of personal contact with anonymous strangers than to the simple inconvenience of answering questions.

  Apprehensiveness may also help explain why the proportion of unlisted phone numbers has grown by two-thirds in
the last two decades and why call screening more than tripled from the late 1980s to the late 1990s. Interestingly, the best predictor of the use of call screening is not affluence or even urbanism, but youth. People under forty-five are twice as likely to screen calls as those over sixty-five, who are (as we have already seen) more trusting and more civically inclined.30 Superficially one might respond that technological development enabled all these changes, but those technologies themselves were surely a response to market demand.

  Other signs, too, of the decline of reciprocity (and its close cousin, civility) can be charted statistically. Voluntary returns of mail census forms declined by more than a quarter between 1960 and 1990. In 1990 the lowest rates of return were among young people, African Americans, and those of us detached from community institutions, precisely the groups within the population that are lowest in social trust. Interestingly, alienation from government itself appears to have played virtually no role.31 In effect, those of us who trust our fellow citizens, but not government, continue to cooperate with the census, while those of us who trust the government, but not the “generalized other,” do not. If it seems to us that other people are playing fair and doing their share, we do, too. If not, not. And “not” is the answer that more and more Americans are giving.

  If fair play toward the “generalized other” is less common nowadays, that should show up in interactions among strangers. Driving is one important domain of anonymous public intercourse in which to chart changing patterns of reciprocity. According to a study by the American Automobile Association Foundation for Traffic Safety, “violent aggressive driving” increased more than 50 percent between 1990 and 1996. The head of the National Highway Traffic Safety Administration conjectures that “road rage” (now common enough to have acquired a name) is a factor in twenty-eight thousand deaths per year. Speeding on the open highway has long been tolerated by large majorities of Americans, but during the 1990s tolerance of speeding in town rose sharply. In 1953, 25 percent of Americans told George Gallup that they had driven over eighty-five miles an hour, compared with 49 percent in a similar 1991 Gallup poll. Older Americans are far less likely to think that you can flagrantly violate the law. In the 1991 Gallup poll 54 percent of all drivers under thirty estimated that you could get away with driving ten miles an hour over the speed limit, compared with only 28 percent of drivers fifty and over. By 1997, drivers who reported that other people were driving more aggressively than five years earlier massively outnumbered those who reported an improvement in civility on the road, 74 percent to 3 percent.32 In short, we all know that other drivers are less courteous nowadays, and “we” are, collectively, they.