Social capital refers to bonds between individuals, both in intimate relationships (primary groups) and in voluntary associations (secondary groups) that make it possible for individuals and groups to achieve a variety of goals. Such bonds have also been claimed to have health promoting effects. In this chapter, I review a variety of empirical studies at both levels of analysis and suggest that the results are mixed, much depending upon the context in which such relationships occur.
Social epidemiologists have long been interested in the ways in which social organization and disorganization influence the health of populations. A recent manifestation is the concept of social capital, which the American sociologist James Coleman defined as follows:
It is not a single entity but a variety of different entities, with two elements in common: they all consist of some aspect of social structures, and they facilitate certain actions of actors—whether persons or corporate actors—within the structure. Like other forms of capital, social capital is productive, making possible the achievement of certain ends that in its absence would not be possible. Like physical capital and human capital, social capital is not completely fungible but may be specific to certain activities. A given form of social capital that is valuable in facilitating certain actions may be useless or even harmful for others (p. S98)1.
Coleman went on to say that, “Unlike other forms of capital, social capital inheres in the structure of relations between actors and among actors. It is not lodged either in the actors themselves or in physical implements of production.”
As Coleman’s definition suggests, social capital is used at two levels of analysis, primary and secondary groups. Primary groups are comprised of family, friends and neighbours and are often thought of as forming networks, or personal communities2. Secondary groups include voluntary associations such as fraternal lodges and civic organizations.
At the primary group level, there have been countless studies of the way social integration and social and emotional support affect both morbidity and mortality. At the secondary group level there have also been many studies arguing that membership in voluntary associations increases trust in one’s neighbours and that civic participation has beneficial consequences for the health of the entire community. Robert Putnam, one of the leading authorities on this type of social capital, writes “Social capital refers to features of social organization such as networks, norms, and social trust that facilitate coordination and cooperation for mutual benefit” (p. 67)3. This type of social capital originates in membership in voluntary associations, and in dense networks of such organizations in any community, because it is in such settings that people learn to trust, to reciprocate and to act in concert to demand public services and good government (p. 182)4.
It is fair to say that there has been great enthusiasm for the beneficial health consequences of social capital, at both the primary and secondary group level. In this chapter, however, I should like to argue that the data are mixed. As with other forms of capital, its consequences may be negative as well as positive.
Social capital and civic participation
I begin with a brief description of Alexis de Tocqueville’s observations made during a famous trip to America in 1831–32. Tocqueville was not only an astute and prescient observer, but his comments on the significance of voluntary associations in American life and on the involvement of Americans in civic affairs have been an inspiration for many contemporary advocates of the benefits of social capital. He observed that the men who initially settled the South were adventurers who came without families in search of wealth. They were followed by “a more moral and orderly race of men” who were, however, “nowise above the level of the inferior classes in England.” And of course the introduction of slavery exercised a “prodigious ... influence on the character, the laws, and all the future prospects of the South” (pp. 16–8)5.
These settlers compared unfavourably with those who settled the North, who “belonged to the more independent classes of their native country.” They were “neither rich nor poor.” All were educated and in proportion to their number possessed “a greater mass of intelligence than is to be found in any European nation of our own time.” They were sober, moral, family men who were accompanied by their wives and children. They were, he continued, Puritans, and “Puritanism was not merely a religious doctrine, but it corresponded in many points with the most absolute democratic and republican theories” (pp. 16–9)5.
Other observers noted that southerners had come from the Celtic fringe of the British Isles and brought with them an entirely different culture than the Puritans. It was a culture that was kin-based, and that emphasized the importance of family and personal honour, the spoken over the written word, agrarian over business values, and leisure over enterprise (p. 268)6.
Many of these cultural differences persisted as Northerners and Southerners pushed westward. People from New England tended to remain in the northern tier of states, whereas southerners moved along the southern tier into the American Southwest. One preacher contrasted the Yankee settlers in Ohio who, “in the strong exercise of social inclination, expressing itself in habits of neighborhood, [formed] villages, and live[d] in them, ... to that sequestered and isolated condition, which a Kentuckian, under the name of ‘range,’ considers as one of the desirable circumstance of existence” (pp. 268–9)6.
Reflecting the cultural differences between North and South were the political differences observed by Tocqueville. He wrote:
Townships and a local activity exist in every State; but in no part of the confederation is a township to be met with precisely similar to those of New England. The more we descend towards the South, the less active does the business of the township or parish become; the number of magistrates, of functions, and of rights decreases; the population exercises a less immediate influence on affairs; town-meetings are less frequent, and the subjects of debate less numerous. The power of the elected magistrate is augmented, and that of the elector diminished, whilst the public spirit of the local communities is less awakened and less influential (pp. 77–8)5.
Complicating the story even further is the legacy of slavery, of which Tocqueville himself was well aware. Stanley Engerman, Stephen Haber and Kenneth Sokoloff have compared the economies of various countries in the Western Hemisphere and have argued that factor endowments—land, climate, indigenous populations—shaped the subsequent growth of institutions and the degree of inequality in each country7. Countries in which conditions had made slavery economically advantageous became far more unequal than those in which small-holdings were relatively equally distributed within the population. Elites in highly unequal countries established institutions that reinforced the early inequality and persist into the present. For example, less is spent on education in such countries because the children of the elite have been educated in private schools; financial institutions and easy credit have not been widely available; and the franchise has not been inclusive. And what is the case at the national level, say between Latin America on the one hand and the United States and Canada on the other, is also true, though to a lesser degree, when the American South is compared to other regions of the United States. That is, the institutional legacy of slavery in the American South has been low educational attainment, high income inequality, and until recently lack of access to the polls when compared to the North, the Midwest, and the West.
All of these differences in the civic cultures and institutions implanted in the North and South continue to make a difference right down to the present. Many studies indicate that the South has the lowest levels of social capital (measured as membership in voluntary associations, and in trust of one’s neighbours), the lowest levels of education, the highest levels of income inequality, and the highest levels of homicide and of overall mortality of all the regions of the USA8–13.
At the state level, therefore, the advantage seems to be to the North, where social capital is higher and mortality lower than in the South. At the metropolitan level, however, a paradox emerges. A study of the impact of racial segregation on mortality in metropolitan regions indicated that, even adjusting for latitude and longitude, segregation was associated with an increase in deaths among African-Americans but not among whites14. But it was also observed that the most racially segregated metropolitan regions were in the Northeast and Upper Midwest, the very areas in which the civic culture that so impressed Tocqueville had taken deepest root. The two are not unrelated.
These cities were America’s industrial heartland, and the relatively unskilled jobs available in their factories and mills provided economic opportunities for vast numbers of immigrants. In the 19th century, cities in the Northeast grew by annexation. They were able to acquire territory at the fringes and incorporate them within a larger city. This was to everyone’s advantage. People in the annexed territory were glad to have city services such as water, sewers, paved roads, and trolley lines, and the city acquired an expanding tax base15.
In the late 19th and early 20th centuries, however, as immigrants from Eastern and Southern Europe, and African-Americans from the deep South began flooding these cities, the older inhabitants moved to the suburbs and began erecting barriers that would prohibit annexation without their consent, including exclusionary zoning and amendments to state constitutions. They drew on the tradition of local community government to protect their economic, ethnic and racial exclusivity. As a result, these cities became the most racially segregated in the country, as well as the least elastic. That is, they are cities that have failed to expand their geographic boundaries to capture suburban growth16. Consequently, the metropolitan areas of which they are a part are divided into many different jurisdictions with their own taxing authorities and school systems. The effect has been to create relatively affluent suburbs and poverty-stricken inner cities, populated disproportionately by poor African-Americans. And in the inner cities access to services and jobs is diminished and mortality and morbidity from many different causes are elevated.
The same period that saw the emergence of exclusionary zoning and independent suburbs is also celebrated as the ‘golden age of fraternalism’ by believers in the efficacy of social capital and American associationalism (when as many of the affairs of society as possible are managed by voluntary and democratically self-governing associations). The burgeoning of fraternal associations in the late 19th century was to a very large degree fuelled by the immigration of people from many different cultures, for mutual aid associations based upon religion and ethnicity were a way to provide burial benefits and sickness insurance to newcomers who would otherwise have been unprotected. They were also among the vehicles by which native-born white Americans could shelter themselves from contact with immigrants. In their ethnic and religious exclusivity, they helped perpetuate and exacerbate cleavages among groups, and they also became an important force resisting the first attempts to create compulsory state sponsored health insurance in the first two decades of the 20th century17,18.
Fraternal organizations, while an answer to one set of serious problems, helped to create obstacles to the provision of universally accessible health care for all Americans. They were not, of course, the only sort of voluntary association to grow during the period between the Civil War and World War I. Settlement houses, playground and parks associations, and hundreds of other organizations devoted to civic improvement and environmental protection emerged at the same time, and with very different goals than the fraternal organizations19. That is precisely the point. The civic culture and rich associational life that have so impressed observers of late 19th and early 20th century America had mixed results. One result was to strangle solutions to problems never anticipated when the colonies were established or when fraternal organizations arose as the answer to the problems of recently arrived immigrants.
Thus when Robert Putnam suggests, “[A]s a rough rule of thumb, if you belong to no groups but decide to join one, you cut your risk of dying over the next year in half. If you smoke and belong to no groups, it’s a toss-up statistically whether you should stop smoking or start joining,” he is using clever rhetoric but giving bad advice (p. 331)20. This suggests confusion between relative risk and attributable risk. More importantly, though, the mixed and uncertain consequences of group membership and social networks, and the unequivocal ill-effects of smoking mean that the decision is not a toss-up, either statistically or for an individual smoker, and it is irresponsible to suggest otherwise.
To recognize the complex nature of civic culture and secondary associations requires more than ecological correlations of measures of trust and mortality. It requires that we think about the history and culture of specific peoples and places21. We see something similar when we turn to a consideration of primary groups and personal communities.
Social capital and primary groups
Just as social capital is fast becoming a paradigm with which to explain inequalities in health at the societal level, so in the form of social support and integration has it become important in studies of the health of individuals22–34. And as in studies at the societal level, so too at the individual level are there results that should encourage a more nuanced way of thinking about the processes at work.
The studies of social support and integration (i.e. involvement with friends, kin and neighbours) on the one hand and subsequent mortality on the other have generally shown that when support is high, the risk of mortality tends to be reduced, even after controlling for various measures of health status. The effects are particularly strong for white males. They are much weaker or non-existent for women and for non-whites, including African-Americans35, Japanese-Americans36 and Navajo Indians37, primarily in rural populations. One commentator has suggested that such anomalies “have to do with the extent to which women and blacks are deeply integrated into stable, rural communities ... Under such circumstances, if these groups routinely obtain large doses of social support, specific measures tapping frequency of contact and size of network may be less differentiating than in less well-integrated communities. Furthermore, these sources of support may be so much a part of the normal daily lives of these people that they commonly go unnoticed and, as a result, are underreported in surveys of this sort. In urban areas, where people are more mobile, their awareness of sources of support and frequency of contact may be heightened”38.
It may well be that the differences are simply the result of methodological difficulties, and that the instruments that have been used to measure integration and support are too blunt to dissect subtle distinctions among populations where support is generally high. On the other hand, the notion that these anomalous results may be explained by the pervasiveness of social support rather than by its absence assumes that which must be demonstrated. It may be that there are true differences in the structure, meaning and functioning of social networks in different populations and settings. For example, in a study in one southern community it was found that African-Americans had lower levels of support than whites39, rather than higher levels as suggested in the previous quotation. Moreover, whereas family networks have been shown to be important sources of support among African-Americans, the conflictual and non-supportive side of such networks has also been noted (pp. 262–3)40.
Furthermore in a variety of studies, women are found to report higher use of support than men, but they are also more likely than men to provide support. “Thus, women may be more likely to be exposed to negative social outcomes regarding both themselves and others than is true for men. Women with larger networks may be more involved in dealing with the stresses of others and thus experience more stress than women with smaller networks or than men” (p. 553)41. Indeed, studies of “care-taker stress” consistently show that people who care for disabled kin are more susceptible to a variety of illnesses than people who do not42–46. More generally, “negative aspects of close relationships [have] adverse effects” on various aspects of health functioning of men and especially of women47.
Finally, living in poverty, living in a rural area, or living in a poor country means that informal social networks and primary groups are called upon to provide a broader range of services than they are in settings where the availability of, and access to, formal services are greater (pp. 583–4)48. Without an infrastructure of functioning public health services, widely accessible health care, public schools, a legal justice system, and other formal institutions, the establishment and maintenance of personal networks are a matter of survival49. But social networks may also impose heavy burdens, for example when reciprocity is required by other members, when gossip and other forms of coercion are used to control behaviour, or when no social services are available to ease the burdens of family care-givers. Kinship and friendship networks are not unambiguously supportive and may exact very high costs.
Not only are different contexts associated with differences in the services provided by personal networks, but the structure of networks is different as well. In general, “studies in several Western countries show that close relatives, especially parents, continue to be of central importance in personal networks. There are, however, differences between rural and urban populations and between social strata in the importance of extended kin compared to friends and acquaintances. In larger cities as well as in middle classes, relations with extended kin become looser, whereas those with friends and acquaintances gain in importance” (p. 104)50,51.
In addition, network structure has profound consequences, even among social strata in the same society. For example, in secure settings such as those in which the middle and upper classes in developed countries live, loose networks based upon friendship are particularly well suited to obtaining jobs as well as needed services from a broad array of formal organizations52. In contrast, dense networks in urban ghettos in the United States isolate people from information about the availability of jobs and services elsewhere in the city53.
Some of the complexities of the associations between social context and the structure, function and effectiveness of social networks are suggested by the results of several studies among Navajo Indians in the American Southwest. Indian reservations like the Navajo’s are generally very rural and characterized by high rates of poverty and unemployment. Traditionally extended families have been both a common as well as the ideal form of social organization and in many instances still are because they are adaptive to the prevailing unstable economic conditions (pp. 243–5)54, (pp. 53–4)55. This pattern is changing substantially, however, as people move to towns on and off the reservation.
In a study of social integration and subsequent mortality among Navajo men and women 65 years of age and above, my colleagues and I found that only marital status was significantly associated with an increased risk of death, and only among men over the age of 75. No other measures of isolation or integration were significant56. This was because, in the Navajo context, social integration meant that the family had to carry a heavy burden. For instance, a significant number of our elderly informants lived with adult children with physical or psychological disabilities for whom no institutional or day care services were available, and for whom the elderly parent was a major source of financial and other support. Rates of depression were significantly higher among these respondents than among people without such burdensome responsibilities.
We also observed in the 1970s that among women of childbearing age, use of contraception was more likely by those who were not living with their kin than among those who were, even adjusting for parity and education. This was because in a generally pronatalist setting, women of childbearing age who lived in extended family arrangements were subjected to intense social pressure, and sometimes physical coercion, from their mothers and husbands not to use contraception, whereas those living neolocally (i.e. in independent nuclear households newly created at marriage) were more autonomous and able to exercise choice57.
But the freedom associated with neolocality and nuclear families is a mixed blessing. Young people who grew up in extended families learned to use alcohol in settings that emphasized responsibilities to kin and therefore shaped their careers as drinkers. Whereas heavy drinking often led to serious difficulties, it was generally outgrown before catastrophe occurred58. However, as the livestock economy has declined, people have left rural areas and residence within extended families and moved to reservation towns where they live in nuclear households, near unrelated neighbours, and attend school with non-relatives. In these new settings, the cross-generational fabric has been torn and a youth culture has emerged which, at the extremes, is characterized by gangs, violence and highly risky drinking59,60. As a result, the homicide rate has tripled since the 1960s. Thus the same freedom from kin that has increased the likelihood of contraceptive use has also resulted in the creation of social networks that encourage destructive and self-destructive behaviour.
These brief examples have been meant to suggest some of the ways in which the larger context may influence the structure, functions and effectiveness of social networks, and how these may in turn influence various dimensions of peoples’ health and use of services. That is to say, social relations are not always supportive and may be damaging, particularly when poverty, unemployment, insecurity and inadequate infrastructure of formal organizations are prevalent. Under such conditions, people have little choice in those upon whom they must depend, and the consequences of enforced dependence on kinsmen may be quite mixed, for they may be oppressive as well as supportive, a hindrance as much as a help. This is unlike the situation of the well-to-do in developed countries, where a good deal of freedom may be exercised in the choice of network members, where networks are looser, and where in any event the need for support from informal sources is less than in many other parts of the world. It is this sort of complexity that explains why social relations which are often assumed to be a form of social capital and of positive value may be of little value or actually harmful.
Some of the most important themes running through the social thought of the past almost two centuries—community, alienation, authority—have represented attempts to come to terms with the consequences of two revolutions, the democratic revolution in France and the Industrial Revolution which began in England61. For some, like Herbert Spenser and William Graham Sumner, the freeing of individuals from the bonds of community and kin, the evolution of nuclear families, and the emergence of laissez-faire capitalism were at once welcome and inevitable62,63. Others were at least ambivalent if not hostile. A widely shared view of thinkers as different as Engels and Marx (p. 324)64 and Tocqueville was that the two revolutions had destroyed aristocratic institutions and local communities which had mediated between individuals and the state.
The result was mass society, alienation and overwhelmingly powerful central governments with no intermediating institutions between them and their citizens. Social capital shares in this tradition, for the answer to the individualism, atomism and alienation of contemporary society—civic cultures woven together by networks of voluntary associations—is not so different from the answers given by many of our 19th century intellectual ancestors.
Thus when Tocqueville went to America, it was to study the most democratic nation of his day. What he found, especially in New England, was municipal governments that were just the sort of intermediating institutions whose absence he feared would lead to despotism in post-revolutionary France. He thought other forms of association could also serve the same function as aristocracy had before the revolution. He wrote:
An association for political, commercial, or manufacturing purposes, or even for those of science and literature, is a powerful and enlightened member of the community, which cannot be disposed of at pleasure, or oppressed without remonstrance: and which, by defending its own rights against encroachments of the government, saves the common liberties of the country (pp. 387–8)65.
The famous late 19th century French sociologist, Emile Durkheim, whose name and study of suicide are cited and recited like a mantra in many articles in the field of social epidemiology, made similar observations. The two forms of suicide that he thought were most important in his time were egoistic and anomic. The egoistic type was the result of loss of cohesion, particularly in religious society (pp. 169, 209)66. That is why he claimed that rates among Protestants were higher than among Catholics. Anomic suicide was the result of loss of restraint, prompted by the growth of industrial society (pp. 254–5)66. Rates and patterns of suicide were, he argued, just one instance of “the whole of our historical development,” the “chief characteristic [of which] is to have swept cleanly away all the older forms of social organization.” He continued, “The great change brought about by the French Revolution was precisely to carry this leveling to a point hitherto unknown ... Only one collective form survived the tempest: the State ... [I]ndividuals are no longer subject to any other collective control but the State, so it is the sole organized collectivity” (pp. 388–9)66. He used the same metaphor of mass society used by many of his contemporaries: “Individuals,” he said, “tumble over one another like so many liquid molecules” (pp. 388–9)66.
We have inherited the world of Tocqueville and Durkheim, but precisely because their insights are so powerful and their writing so persuasive, we run the risk of allowing them to over-determine how we understand ours. Durkheim’s study of suicide, for instance, has been so influential and is so frequently cited, but not because his analyses were valid. Indeed, they probably were not67,68. It is influential because he addressed concerns that we share and provided answers to which we resonate. The same is true of our response to Tocqueville.
But our world is different from theirs. Local governments may involve their citizens in civic life, as Tocqueville suggested, but they may also promote exclusivity and protect citizens from integration into larger regional structures that may be widely beneficial. Associations “for political, commercial, or manufacturing purposes” may protect their own interests, as Tocqueville and Durkheim both suggested, but in the process may be harmful to the health and well-being of the public. This is amply demonstrated by the opposition of the tobacco industry trade association to limitations on cigarette advertising and sales, and by the destruction of President Clinton’s plan for health care reform by a coalition of voluntary associations including the National Rifle Association, the Christian Coalition, the National Federation of Independent Businesses and the Health Insurance Association of America69.
Churches may be the most integrative institutions in Hispanic and African-American communities, but their deep conservatism prevented them from acting early in the AIDS crisis and contributed to its severity70. And kinship and friendship networks may provide support to their members but may also exact high costs, resulting in resentment and stress as well as delinquency and criminal behaviour. Social capital may be valuable in facilitating certain actions, as James Coleman wrote, but “may be useless or even harmful for others (p. S98)”1.
The quest for community has been a recurring theme in social and political thought. Certainly in the United States, perhaps the most individualistic of nations, it has been significant from the 19th century to the present in the form of utopian communities and encounter and support groups. Such communities generally represent a withdrawal from civil society rather than engagement with it, and for this reason, the new republican theorists distance themselves from communitarianism. Nonetheless, they share common concerns and insights. Most importantly, both abjure laissez-faire individualism in favour of the beneficial effects of association. I have argued that whatever other benefits may derive from association—and I believe them to be both real and significant—improved health is not invariably one of them, for primary and secondary group ties may bind us together, but they may also imprison and divide us.
This is a revised version of the following chapter: Kunitz SJ. Accounts of social capital. In: Leon DA and Walt E (eds), Poverty, Inequality and Health: An International Perspective. Oxford: Oxford University Press, 2001. Reproduced by permission of Oxford University Press.