One of the main concerns is the lack of clarity and conceptual confusion that surrounds much of the work about social capital. Even the enthusiasm expressed by the Performance and Innovations Unit was tempered by the plethora of definitions encountered in their literature review, leading them to caution about the potential mis-specification or ambiguity in some of the models used in the empirical literature (PIU 2002). Opinions are divided about whether this is because social capital is a falsely inflated concept (Labonte 1999) or because of the lack of clear guiding principles underpinned by a sound theoretical base (Lomas 1998, Wall et al 1998). Whatever the reason, unclear concepts create problems for research.

Kawachi et al (2004), for example, identified 31 studies from across the world, which illustrate the wide variability of definitions and unclear concepts. The results showed that the presence of social capital led to better health outcomes in the large majority of studies, but the indicators used varied widely (Box 5.2). The authors highlight a key difference: that some measures (mortality and crime rates, for example) relate to collectives or whole populations, while others relate to aggregated measures (such as self-rated health and self-esteem) drawn from individual respondents. This raises the question about who ‘owns’ social capital, and whether it is legitimate to measure it in a community, but apply it to individuals, or vice versa. Without theoretical and conceptual clarity about exactly what is being measured or why, it is hard to have faith in the results, and Hean et al (2003), among others, have called upon researchers to explain the mechanisms being measured in theoretical terms.

Concerns about a lack of clarity are linked with the second major area of criticism, which can be summarized as a lack of theoretical awareness. Critics suggest that much of the social capital literature is merely rewriting well-established theories; social capital is described as ‘old wine in new bottles’ or as a ‘catch-all phrase’ for social support and democratic theory (Berkman et al 2000, Edwards and Foley 1998, Labonte 1999, Muntaner 2004, Portes 1998). Such writers point out that we have known the importance of social circumstances on the quality of life and health since the work of 19th-century sociologist, Emile Durkheim. The collective impacts of social relationships and social trust on health have been very well documented, and there is an expansive empirical literature about both social networks and the different forms of community, including its cohesion. Social capital literature is viewed as naive and uninformed by these critics, who suggest it fails to take sufficient account of existing social theories. Szreter and Woolcock (2004) suggest that the majority of the literature about social capital and health can be divided under two headings: that which is concerned with how psychosocial mechanisms, such as social support, affect health and that which purports to explain health inequalities (as outlined above). These two large fields of interest were well represented in psychological and sociological literature long before social capital emerged as a concept of interest.

In exploring psychosocial mechanisms, Stewart-Brown and Shaw (2004) carried out a study that differed from those taking the community as a starting point, instead bringing the family into sharper focus. They hypothesized that the roots of social capital were embedded in relationships in the home during childhood, which would influence health in later life. By comparing the results of two systematic reviews with published birth cohort data across three cohorts (born in 1946, 1958 and 1970), they were able to provide strong evidence that the quality of relationships in the home in childhood has an impact on both physical and (particularly) mental health in adulthood, regardless of socio-economic circumstances. This important analysis identifies issues such as warmth, relationships, social support and the developmental importance of children growing up in circumstances that are conducive to trust and networking, so that as adults they might both accept trust and be the kind of person, or neighbour, who is trusted by others.

Despite the value of the work, reflecting the criticisms outlined above, Stewart-Brown and Shaw’s (2004) study of social trust and relationships would have been equally useful if the term ‘social capital’ had not been used. However, the study does underscore the potential for social capital to be both generated and used within relationships, which is the basis for the norms of reciprocity and obligation that arise in much of the social capital literature. In terms of measurement, it is important to know whether data are being drawn from a social capital user or generator, or from both (Hean et al 2003). This distinction is similar to one suggested by Snijders (1999) to distinguish between different social capital pathways. He suggested using the term ‘complementary’ to describe when one person has the social capacity to offer some element, such as support or information, from which another can benefit; and ‘symmetrical’ for when there is a collaborative or cooperative effort from which all parties can benefit equally. The distinction is important for researchers, so they know what to measure. In addition, it becomes important for practitioners and those interested in improving social capital in an area. Is it possible for an outside person, such as a health professional, to enable the generation of social capital through networks and groups, for example? Community development literature is divided on whether the activities of an insider or outsider, or lay or professional worker, are most effective; the social capital literature has not really engaged with these issues.

Both a lack of conceptual clarity and the failure to specify the theoretical basis for social capital may be regarded as acts of omission. They are criticisms that have largely been taken on board by researchers in the field of social capital, with varying degrees of success. The third group of complaints represents more fundamental differences of opinion. Kanazawa (2006) is something of an outlier in the debates. He draws on evolutionary genetics to suggest that average intelligence quotient (IQ) as measured across countries provides a better explanation than social capital for international health inequalities, an idea that is given short shrift by Wilkinson and Pickett (2007), who reject the way he conceptualizes IQ.