A mericans now understand that their health is at risk if they smoke, overeat, and fail to exercise. But a growing body of evidence suggests that public health also depends on a less widely understood influence-social cohesion. And while many Americans have stopped smoking, gone on diets, and put on jogging shoes, American society has become, if anything, less cohesive.
Consider what happened in Roseto, a small Italian-American community in eastern Pennsylvania. During the 1950s, when the town first caught the attention of medical researchers Stewart Wolf and J.G. Bruhn, Roseto posed something of a mystery. Death rates in the small town of about 1,600 people were substantially lower than in neighboring communities. In particular, the rate of heart attacks was about 40 percent lower than expected and could not be explained by the prevalence of factors known to increase the risk of the disease. Citizens of Roseto smoked at the same rate as neighboring towns, they were just as overweight and sedentary, and their diet consisted of about the same amount of animal fat. But the one feature that stood out was the close-knit relations among residents in the community. The town had been originally settled by immigrants during the 1880s, who all came from the same village in rural Italy. The researchers noticed the social cohesiveness and ethos of egalitarianism that characterized the community:
Proper behavior by those Rosetans who have achieved material wealth or occupational prestige requires attention to the delicate balance between ostentation and reserve, ambition and restraint, modesty and dignity. . . . The local priest emphasized that when preoccupation with earning money exceeded the unmarked boundary it became a basis for social rejection. . . . Rosetan culture thus provided a set of checks and balances to ensure that neither success nor failure got out of hand. . . . During the first five years of our study it was difficult to distinguish, on the basis of dress or behavior, the wealthy from the impecunious in Roseto. . . . Despite the affluence of many, there was no atmosphere of "keeping up with the Joneses" in Roseto.
But as young people began to move away to seek jobs in neighboring towns and the community entered the mainstream of American life, the social taboos against conspicuous consumption began to weaken, as did the community bonds that once maintained the town's egalitarian values. About a decade into the study, the researchers noted:
For many years the more affluent Rosetans restrained their inclination toward material indulgence and maintained in their town the image of a relatively classless society. When a few began to display their wealth, however, many others followed. By 1965 families had begun to join country clubs, drive expensive automobiles, take luxury cruises, and make flights to Las Vegas.
The unforeseen consequence of improved material well-being and, probably more important, rising socioeconomic disparities was that the incidence of heart attack in Roseto caught up with neighboring towns within a span of a decade.
The notion that social cohesion is related to the health of a population is hardly new. One hundred years ago, Emile Durkheim demonstrated that suicide rates were higher among populations that were less cohesive. In 1979, after a nine-year study of 6,928 adults living in Alameda County, California, epidemiologists Lisa Berkman and S. Leonard Syme reported that people with few social ties were two to three times more likely to die of all causes than were those with more extensive contacts. This relationship persisted even after controlling for such characteristics as age and health practices, including cigarette smoking, drinking, exercise, and the use of medical services. The basic findings of the Alameda County Study have since been confirmed in more than a half dozen epidemiological studies in different communities.
These findings have ominous implications if the political scientist Robert Putnam is right that social capital is declining in America [see "The Strange Disappearance of Civic America," TAP, Winter 1996]. Putnam's memorable metaphor for this change is bowling league membership, which has declined while bowling overall has increased. By social capital Putnam means the invisible glue that holds society together-the social networks, norms, and trust that enable groups of individuals to cooperate in pursuing shared objectives. On the basis of research in Italy and elsewhere, Putnam argues that social capital is a major contributing factor in economic growth [see "The Prosperous Community: Social Capital and Public Life," TAP, Spring 1993]. In fact, as the public health research shows, the harm from weakening social cohesion may not only be civic and economic-it may also be physical.
IT DOES HURT TO BE ALONE
To explore this question, we set out to test the relationship between social capital and public health at the state level. In fact, there are quite marked geographical variations in civic trust and association membership across the United States, and when these indicators of social capital are arrayed against regional differences in mortality and morbidity, the resulting correlations are striking. The chart "Social Capital and Mortality Rates" (below) shows the relationship between the level of civic trust and the age-adjusted rate of death from all causes for the 39 states for which data were available in the National Opinion Research Center's General Social Surveys. The lower the trust between citizens-as indicated by the proportion of respondents in each state who believed that "most people cannot be trusted"-the higher is the average mortality rate.
A similar relationship with mortality prevails for the per capita membership of state residents in voluntary associations. These relationships between social cohesion and mortality hold among both whites and African Americans, as well as among men and women, and they persist after statistical adjustment for state variations in median household income and proportion of households living below the federal poverty threshold.
The figure below, "Social Trust and Quality of Life," displays the correlation between level of civic trust and a measure of self-reported well-being. The National Center for Chronic Disease Prevention and Health Promotion employed the Behavioral Risk Factor Surveillance System (BRFSS) to ascertain the proportion of residents in each state reporting that their health was only fair or poor as opposed to good or excellent. (The BRFSS is a representative, random telephone survey that sampled more than 350,000 community-dwelling American adults between 1993 and 1996.) Again, there is a striking correlation between social capital and quality of life.
But does "bowling alone" really increase the likelihood that you'll get sick? Putnam's reference to the decline in bowling leagues evinced skepticism from some critics. Katha Pollitt, for example, pointed out that the popularity of bowling leagues emerged from a particular period in American blue-collar culture that permitted husbands plenty of boys' nights out (think of the memorable first glimpses of Marlon Brando in A Streetcar Named Desire). Other critics have pointed out that declining bowling league memberships may be offset by increased participation of other kinds, such as coaching and playing in youth soccer leagues.
Nonetheless, bowling league membership turns out to correlate rather well with who lives or dies (see "Bowling League Membership and Mortality," below). To paraphrase John Donne, no man or woman is an island entire of itself-therefore we should never send to ask for whom the ball rolls.
INEQUALITY AND PUBLIC HEALTH
Another feature of a society that may influence both its cohesiveness and its members' health is the level of economic inequality. In many countries, notably America, income and wealth are becoming more concentrated. According to a Census Bureau report released last year, the share of total income going to the top fifth of American households increased from 40.5 percent to 46.9 percent between 1968 and 1994. By contrast, the shares of the bottom 80 percent either declined or stagnated. The biggest income gains went to the top 5 percent of households, whose share of the economic pie increased from 16.6 percent to 21 percent. In 1994, the average income among the top 5 percent of households was more than 19 times that of the bottom 20 percent.
Might this polarization of incomes be loosening the social cement? In a forthcoming article in the American Journal of Public Health, we argue that this is the case. "Income Inequality (Robin Hood Index) and Social Trust" (below) shows the rising trend in income inequality plotted against the steady decline in civic trust, as tracked by the General Social Surveys. The measure of income inequality we used is the Robin Hood Index, which equals the proportion of aggregate income that would have to be redistributed from households with disproportionate earnings to those earning less, if incomes were to be level. The higher the Robin Hood Index, the bigger the income gap. As "Income Inequality and Social Trust" shows, the larger the income gap, the lower is citizens' trust in each other. Nearly identical results are obtained when we plot income disparity against per capita participation in voluntary associations.
Comparing public health and income distribution across countries lends further credence to the notion that income distribution plays a greater role in the quality of public health than more traditional indices do. In his recent book, Unhealthy Societies: The Afflictions of Inequality, economic historian Richard Wilkinson argues forcefully that the life expectancy in developed countries cannot be explained by differences in their absolute standard of living as measured, for example, by per capita income. Rather, a population's health depends more on the level of economic inequality.
The United States, despite having one of the highest living standards in the world (the real gross domestic product [GDP] per capita was $24,680 in 1993), has a lower life expectancy (76.1 years in 1993) than less affluent but more egalitarian countries like the Netherlands (GDP, $17,340; life expectancy, 77.5 years); Israel (GDP, $15,130; life expectancy, 76.6 years); or Spain (GDP, $13,660; life expectancy, 77.7 years). In fact, societies with the smallest income differences between rich and poor, such as Sweden and Japan, tend to enjoy the highest life expectancy (78.3 and 79.6 years, respectively). An egalitarian distribution of wealth and income seems to imply a more cohesive, harmonious society. The quality of social relations, Wilkinson concludes, is the prime determinant of a country's human welfare and quality of life.
What does this imply for our future quality of life in this country? Is what has been happening to American society simply a case of Roseto writ large? Two studies published simultaneously in the April 20, 1996, issue of the British Medical Journal, including one we conducted, found that differences in income distribution across the 50 states were highly correlated with mortality rates, including deaths from heart disease, homicides, and infant mortality. To be sure, overall life expectancy in the United States has been steadily improving due to advances in medical treatment and the prevention of disease through lifestyle changes. But mortality might have declined more if income inequality had not risen. Our model suggests that for every percent increase in income inequality, the overall death rate is 2 to 3 percent higher than it needed to have been. By any definition, this is an important public health problem.
In recent years, unfortunately, government policy has tended to reinforce growing inequality, which is unsurprising in view of the disproportionate political weight that the well-off carry [see Sidney Verba, Kay Lehman Schlozman, and Henry E. Brady, "The Big Tilt: Participatory Inequality in America," TAP, May-June 1997]. The danger is a self-perpetuating cycle of growing income inequality, growing political inequality, and diminishing social capital. And because health too is at stake, it is no exaggeration to say that breaking that cycle will affect the body politic in every sense.
