Putting social capital in its place.

I. Introduction

'[If] people cannot trust each other or work together, then improving the material conditions of life is an uphill battle' (Evans, 1997a: 2). Recently, trust, cooperation and other similar processes have been brought together under the concept of social capital. More specifically, social capital refers to norms of trust and reciprocity and to networks, associations and organisations that constitute social resources for individuals, and which facilitate collective action for mutual benefit (Woolcock, 1998). Strictly speaking, these social resources are not capital, and they are referred to as capital here solely in a metaphorical sense, since capital proper is an exploitative relationship between capital and labour, and resources only become 'capital' in this relationship.

The popularity of the social-capital concept partly reflects the growing realisation among non-Marxist social scientists--for Marxists, economic processes are always social processes--that economic processes are linked to social relations, which, in turn, influence these processes (Granovetter, 1985). It also reflects the current worldwide neoliberal agenda. In particular, social-capital research on how non-market processes such as the state, trust and customs grease the wheels of market fits in well with the neoliberal agenda of making (imperfect) markets more efficient. Underneath the ever-growing popularity of social-capital literature there is a major problem, however. This literature, in general, tends to under-stress the class character of social capital. I am not suggesting that the class character of social capital has been totally neglected. Indeed, a few scholars do attempt to incorporate the class dimension (Bourdieu, 1986; Duncan, 2001; Portes & Sensenbrenner, 1993), and this paper seeks to extend and contribute to that work.

The paper has six sections, including the introduction. Sections 2 and 3 discuss 'society-centric' and 'state-society relation' approaches to social capital, respectively. Section 4 presents a class-based critique of these approaches. In section 5, I present a political-economy approach to social capital, examining the different ways in which the class context of working-class social capital enables and constrains its production in particular places and over space. In the final section, I draw out some of the political implications of my approach.

2. The society-centric approach to social capital

In much of the literature on social capital, it is seen as inhering in civil society, but outside the state. Social capital, in this approach, is said to be a non-market means of addressing imperfections or failures of the market. It plays this role by promoting information-sharing through networks and lowering transaction costs, by discouraging opportunistic behaviour, and by helping collective action to (for example) manage common property (Serageldin & Grootaert, 2000:47 49; Mearns, 1996).

There have been several quantitative studies on social capital. These claim to show that social capital (trust, civic norms, etc. in civil society) is associated positively with income and income equality--see Knack and Keefer's (1997) study of a sample of twenty-nine market economies--and with poverty reduction in Indian states, for which see Morris (1998). Social capital is associated negatively with infant mortality and income inequality in thirty-nine US states (Kawachi et al., 1997). World Bank researchers assert that the social capital of rural households--social capital understood as networks and associations--is positively and causally associated with household expenditures and welfare in African and Latin American countries (Grootaert et al., 2002; Grootaert & Narayan, 2004). But note that these studies tend to equate correlation with causation. It is also possible that when social capital is treated as one factor operating along with many others--and especially with class-related factors such as the strength of working-class power--it may have weak association with its purported social outcomes. A study of health and social capital in sixteen wealthy countries suggests this to be the case (Muntaner et al., 2002). Besides, there is also the possibility that both social capital and its supposed effects (e.g. economic growth) are a result of a third factor, such as state intervention or specific class relations.

Along with the quantitative studies, several case studies on social capital have also appeared. Putnam's (1993) Italy study claims that greater degrees of citizens' civic engagement breed trust and cooperation. These, in turn, discourage people's 'free-rider' attitudes and thus resolve the collective-action problem involved in, for example, monitoring government performance. This makes governments become more efficient and more democratic in delivering public goods, causing greater well-being. Kennedy et al.'s (1998) study finds that in Russia, those who lack social capital in the form of support from friends and family are more likely to be vulnerable to economic hardship caused by the transition to a market economy.

Social capital in the form of entrepreneurial networks is shown to have enhanced the performance of the manufacturing sector by facilitating flows of information between enterprises in Ghana (Barr, 2000) and in Nigeria (Brautigam, 1997). Uphoff and Wijayarama (2000) report social-capital formation in an irrigation scheme in Sri Lanka, in which ethnic cooperation was demonstrated by upstream Sinhalese farmers sharing water with downstream Tamil farmers (p. 1880). Cooperation among farmers reduced water wastage by individual farmers, who earlier had no concern for people lacking water (Uphoff & Wijayaratna, 2000).

These and many other similar studies shed light on the social capital of civil society, i.e. relations of trust and reciprocity among friends and neighbours, and their local networks and organisations. However, these studies not only under-stress the class dimension of social capital (more about this later), but also generally have a society-centric view of it. That is, they tend to overlook the possibility that relations between civil society and state officials, and not just relations within civil society, can constitute social capital. The state is seen as being merely influenced by social capital (Levi, 1996: 50; Booth & Richard, 1998; Unger, 1998). Arguably, if social capital is about social relations of trust and cooperation, these relations can potentially include the relations between the state and society. But social capital is not explicitly conceptualised as such, and some scholars even think of the state as antithetical to social capital (Chhibber, 2000:299).

3. The state--society relation approach to social capital

Some recent works attempt to counter this under-conceptualisation of the state. In these, social capital is seen as inhering 'not just in civil society, but in an enduring set of relationships that spans the public-private divide' (Evans, 1997b: 184). So, apart from civil society, another locus of social capital is the 'zone' of interaction between civil society and the state. The state in question is a relatively autonomous, competent state with coherence and credibility. It is sufficiently embedded in civil society. It has a hierarchical division of labour that allows communities and officials to work out their problems flexibly at the local level, without undue interference from above (Evans, 1997b: 194).

Social relations between the relatively autonomous state (officials) and society can be characterised as synergy. Synergy, following Evans and pace Woolcock (1998), has two aspects: embeddedness and complementarity. 'Embeddedness' refers to (face-to-face) ties of trust between citizens and public officials, who are 'more thoroughly part of the communities in which they work' (Evans, 1997b: 184). Examples of embeddedness include joint business-government deliberative councils, and neighbourhood meetings in which officials and communities participate in order to resolve conflicts. 'Complementarity' refers to the fact that the state provides things that the communities do not or cannot have, but which complement what they do have (Evans, 1997b). The scale and bureaucratic organisation of the state allow it to more effectively provide certain kinds of collective goods that complement the inputs more efficiently delivered by private actors. These can be material goods such as the physical infrastructure, including means of transportation and communication, which contribute to social connectedness. As a complement to state provision of these goods, citizens contribute local knowledge and experience.

This broader notion of social capital is exemplified in several empirical studies. Ostrom (1996) shows that poor neighbourhoods in a Brazilian city had to rely mainly on the government to produce trunk sewer lines, which they could not have provided on their own. Face-to-face interaction between them and officials, however, facilitated the maintenance of the system. Similarly, joint forest management by local communities and state officials is said to have increased village incomes in India's state of Gujarat (Serageldin & Grootaert, 2000). A study by two geographers, Bebbington and Perreault (1999), shows that state officials with a bias towards indigenous highland communities in Ecuador formed social capital for these communities by helping them build local organisations, and by linking these to organisations at higher geographical scales. These organisations have in turn led to state policies benefiting the highlanders. Tendler (1997) investigates how in Brazil's Ceara state, the government invested in building trust relations between officials and people, which increased interaction between them and improved the performance of a public health campaign, reducing infant mortality.

These studies show that social capital is formed through close interaction between citizens and state officials. 'Norms of co-operation and networks of civic engagement among ordinary citizens can be promoted by public agencies' through their synergy with civil society (Evans, 1997b: 178, emphasis added; see also p. 184), as can social capital in the form of trust.

4. A class critique of the social capital-literature

Both of the above approaches to social capital--and especially the former, with its link to rational-choice theory--under-stress the fact that society, communities and places are class-divided; that so-called ordinary citizens belong to particular classes; and that the most important context in which social capital is produced is the class context. The idea that social capital is constituted by social relations (Woolcock, 1998) is a useful one. But the social relations that social-capital experts talk about are, generally speaking, not (explicitly) conceptualised as class relations. (Neither are these relations conceptualised specifically as relations of race or gender, and here, I am abstracting from these.) These are much rather seen as relations between individuals, or aggregates of individuals. Social capital is said to be about 'how individuals achieve co-ordination and overcome collective-action problems' (Ostrom, 2000: 173, italics added). Emphasis on the individual, on the community (a close cousin of social capital) and the local scale on the one hand, and the under-theorisation of class on the other are two sides of the same coin. The relative neglect of class (and especially of a relational-conflictual view of class) is expressed in different forms, as briefly discussed below. (1)

Social-capital scholars emphasise the moral-economy aspect of social capital, i.e. reciprocity and self-help at individual and community/local levels. Certainly, moral economy (especially in the sense of the moral responsibility of the rich towards the poor) as a basis for self-help may be important in specific places or instances. But it should not be considered in isolation from political economy (i.e. the unequal distribution of material resources between classes and the attendant issues of power). That moral economy or mutual assistance can mean 'narrow and often risky personalized dependencies of people on each other' (Levi, 1996: 51), and especially of the poor on the rich, is not systematically recognised.

A few scholars have commented on the neglect of political power in social-capital literature (Foley & Edwards, 1996). But they tend to overlook the class-specific implications of this neglect. In particular, it is problematic not to specifically look at the social capital of the working class. For one thing, the working class (in both rural and urban spaces) has a unique social-political resource in its hands. It is the class that can stop capitalist production and valorisation; it can take on the capitalist class in a way that no other class or group can (Wood, 1998). Members of the working class can share the ways and the means of exercising this unique power/ resource among themselves; and since sharing and reciprocity are important to social capital, this class-specific form of sharing and reciprocity--emphasised in the radical literature on the working class--cannot be neglected. Also worth noting is that the political organisation of the working class is a precondition for its economic survival to a greater extent than it is for the capitalist class, whose power lies relatively less in its actual political organisation (its public political power) and more in its ownership and control over property and the accumulation process (its private political power).

Any neglect of working-class political mobilisation is therefore necessarily biased against the working class. Broadening the scope of the social-capital concept to include the state is not enough to counter the neglect of class in the literature, since the state itself is primarily an institution with class-specific actions and effects. This is something that most of the social-capital scholars, including Evans, tend to ignore, in part because of their Weberian approach to the state.

In addition, to the extent that social-capital writers do deal with issues of power, they tend to have a socially undifferentiated view of associationalism and solidarity, such that there is no recognition that political parties, football clubs and bird-watching groups do not have the same political effects (see Levi, 1996: 49). They emphasise mutually beneficial coalitions across social and economic divisions (World Bank, 2001; Putnam, 1993), which are assumed to benefit society as a whole as if there were no class (and other) conflicts of interest in society.

And, to the extent that class is discussed at all in the social-capital literature, the working class tends to be left out. Much of the literature, including research on industrial districts and on the politics of economic development in geography and development studies, discusses the ways in which capitalists and other propertied groups use relations of trust and networks in order to improve their competitiveness in particular places, and to solve coordination problems. Thus when class appears, it does so in the form of mainly propertied classes, including small capitalists and the petty bourgeoisie. This stress on small property owners can be seen in geographical work on social capital (Bebbington) and small-scale industrialisation, and in sociological work on ethnic capitalism (Portes). This would be less problematic if the propertied class were seen in relation to the property-less. A dialectical-relational view of class is often missing.

It might seem that I am exaggerating the (working) class-blindness of the social-capital literature; and certainly, in the literature, there is a discussion of poor people. But even in this discussion, in which social capital is said to be the 'capital' of the poor (2) (World Bank, 2001), the class character of social capital is not examined. There are at least two reasons for this neglect. For one thing, the class character of the poor themselves is usually left out: the fact that the poor generally belong to classes--wage-earning and semi-proletarian classes--whose material interests are in conflict with the classes having ownership/control over resources (Wright, 1995) tends not be considered. In addition, social-capital scholars often assume that as the social capital of society develops, the poor will benefit--and indeed, that they benefit from social capital more than the non-poor. This assumption is arguably the social-capital version of the neoclassical trickle-down theory of poverty.

Not considering the class character of poverty obviates the need to conceptualise the way social capital develops in, and is constrained by, the class context in which the poor live. There is a discussion, for example, on the social capital of peasants (semi-proletarians) (Bebbington & Perreault, 1999). But discussing peasants' social capital without considering the powers and liabilities of peasants as a class whose interests are in conflict with those of landlords, and the way in which their class position constrains (and enables) their social capital, is problematic. This neglect of class might be understandable were there not a body of literature to connect it to. But there is indeed a large amount of literature on self-help and related activities within working-class communities (Bourke, 1994; Hopkins, 1995; Lee, 2003; Morriss & Irwin, 1992; Short, 1996; Thompson, 1963). One purpose of this paper is to point out how social-capital writers can learn from this literature.

Social-capital enthusiasts' uneasiness with class is very well expressed in the following quote: 'If a community is riven by conflicting interests, the nature and meaning of social capital becomes more complicated' (Evans, 1997b: 196; also see Duncan, 2001: 75). If this is so, why not take conflict and its underlying theoretical underpinnings, including class, as the starting point rather than social capital--and why not discard social capital altogether? That is Ben Fine's suggestion (1998). But I will take a slightly different route, based on what he calls 'a less acceptable alternative' to jettisoning social capital altogether: that social capital be used as an investigative category, and that it be directed to take account of economic power and class context (Fine, 2000). My intention is to develop a class-based, political-economy approach to social capital, which will unpack the opportunities of--and, more importantly, limits to--social capital as they are rooted in class, in specific places and over space. The question I ask is this: to the extent that social capital is a useful category, what is its class character?

5. Towards a political economy approach to social capital

Underlying the political-economy approach to social capital is a specific approach to class and the state (Jessop, 1990; Das, 1996; Camfield, 2004; Wood, 1998). Class is a structure of relations of, and processes (re-)producing, inequality in the distribution of means of production and the resultant relations/processes of exploitation between classes. Class structure defines limits within which certain outcomes are more likely than others (Wright, 1978; Wood, 1998). It is both enabling and constraining (Callinicos, 1988: 235). In terms of the relation between social capital and class, class imposes limits within which social capital is produced and has effects on different classes (and on other relations).

Within the limits of class, certain methods of social-capital production and certain effects of social capital are more likely than others. Social capital is produced and used by particular classes. Social capital is about norms of reciprocity, and networks/organisations that foster these norms. Cultural and political norms always presuppose material resources/ practices (and this is something that most postmodern scholars tend to forget). Without these resources, norms as well as networks/organisations have limited causal powers. And access to material resources is influenced by class relations more than by any other set of social relations.

I will look at the social capital of the working class only, since space constraints do not permit a discussion of the propertied class's social capital. The working class is defined here as the class that is excluded from the ownership of, and control over, means of production, and whose main source of income is the sale of labour power. The working class thus defined can include labourers with small properties, as long as the main class contradiction the working class faces is with capitalists rather than with any other class/group, although here I am specifically dealing with a subset of the working class: poorly paid, unskilled/semi-skilled labourers with limited security of employment, and a very low level of autonomy at work. I use 'working-class social capital' (3) primarily as an investigative category with which to look at the ways in which the working class (re)makes and lives its life (Thompson, 1963) and creates its history and geography (Herod, 1997; Harvey, 1982a).

In struggling for better economic--political conditions in particular places within the limits defined by class structure, workers employ several strategies. One of them is social capital, which they use to resolve their own collective action problems and to build connections with the state and with one another. Working-class social capital, like social capital in general, exists within civil society ('micro level') and in state--society relations ('macro level'--categories used solely for convenience of labelling). Its production is enabled and constrained by processes broadly derived from its class context.

Micro-level working-class social capital is the social capital of the working class inhering in civil society. It refers to norms of trust and reciprocity, and to networks and organisations among working-class people. Working-class micro social capital has class-specific effects.

Norms of trust and reciprocity can discourage free-rider behaviour and facilitate collective actions such as the building of physical and social infrastructure in working-class communities (e.g. community centres, roads, library, reading groups), which can enhance their social connectedness; launching political action against employers paying low wages; or monitoring corrupt officials implementing antipoverty policies. Norms of trust and reciprocity can lead to informal mechanisms of support within working-class communities.

These can take such forms as the informal borrowing of food and other items of necessity, (4) worker-managed eating places, micro-credit groups, and the sharing of information or tacit knowledge between (propertied) labourers--those whose income comes mainly from wages, but who supplement their wage income through petty entrepreneurial activities. Working-class people also use relations based around family and kinship as survival strategies (Argyle, 1994: 79). Self-help (e.g. food sharing; informal credit) is particularly important during strikes, since it can increase workers' staying power.

The production of working-class social capital is enabled and constrained by class processes. First, let us consider the enabling class processes. Useful here is the concept of 'organicity of classes'--a crucial concept in contemporary class theory (Bodeman & Spohn, 1986). It refers to 'the primordial relations, the particular [class-specific] ways of life and culture, the common outlook and the interpersonal ties within which ... [the working class is] embedded' (ibid: 10) (although primordial relations can also create disunity in the working class).

The concept of organicity of classes counters the idea of what Bodeman and Spohn call the 'naked proletarian'--that the working class is without traditions, ties of family, kinship and neighbourhood, etc. As E. P. Thompson aptly stated, the working class is a social and cultural formation, like capitalism itself. Commenting on the influence of religion on class struggle, he referred to processes that arguably connote social capital:

   No ideology is wholly absorbed by its adherents: it breaks
   down in practice in a thousand ways under the criticism
   of impulse and experience: the working class community
   injected into the chapels its own values of mutual aid,
   neighbourliness and solidarity. (1963: 392; italics added)

Indeed, mutual aid and neighbourliness are important forms of social capital. If, as Marx says in his famous preface to A Contribution to the Critique of Political Economy, the material conditions under which people live generally structure their 'social consciousness' and political organisation, then it is not difficult to argue that the common, class-specific ways of life of working-class families can create a specific type of bond between them--what is known as 'bonding social capital' in the social-capital literature, i.e. social relations and norms of reciprocity among people of similar socioeconomic status (Putnam, 2000: 22-23).

   As Marx says in The Eighteenth Brumaire, in order to be
   a politically organised class, workers must have a
   community, a national bond, and a political organisation.
   Community, bond and organisation are also keywords in
   the literature on social capital.

Portes develops a four-fold typology of sources of social capital. One is 'bounded solidarity' (Portes & Sensenbrenner, 1993: 1325). This concept is exemplified by Marx's analysis of the rise of working-class consciousness: 'The weapon of the working class in this struggle is precisely its internal solidarity born out of a common awareness of capitalist exploitation' (ibid: 1324). As Marx (1975:172) says in The Poverty of Philosophy, the domination of capital creates the common situation and common interests of the working class. An emergent sentiment of 'we'-ness among workers who face similar difficult situations develops; this is a sentiment based on moral imperative, rather than on rational calculations of costs and benefits, and promotes solidarity (ibid: 1327-8; Callinicos, 1988:199-203). Indeed, I argue, their common experience of relative (and for many workers, absolute) poverty as well as of insecurity and vulnerability is a potential basis for reciprocity and networks. Portes says that 'If sufficiently strong, this emergent sentiment [of weness] will lead to the observance of norms of mutual support [e.g. exchange of gifts], appropriable by individuals as a resource in their own pursuits' (p. 1325). This is similar to the support that newly arrived ethnic-minority immigrants get from existing ethnic-minority immigrants, who have a sense of solidarity based on their perception that they are discriminated against (ibid.). (5)

Workers' class position not only enables the production of workers' micro social capital; it also constrains it. Given their class position, workers have limited material resources to share among themselves. And without resources, informal rules of sharing in a place (e.g. in a neighbourhood) and over space (e.g. between neighbourhoods) are difficult to sustain for long (in their active form). The social resources of the working class, and its networks and connections, tend to be indeed socially and spatially limited. The working class has 'less' social capital than other classes. This is confirmed by a few empirical studies. In Britain, for example, '[there are] deep differences in the connections that people in different class situations have to the community.... On average, people in the middle class have twice as many organisational affiliations as those in the working class' (Hall, 1999: 438).This is consistent with more recent findings from Britain that people in disadvantaged positions are more likely to draw social capital from weak ties, while those in more advantaged positions are more likely to draw social capital from formal civic engagement and networks (Li et al., 2005).

In addition, a higher proportion of working-class people 'suffer[s] from a complete absence of social support' than do members of other classes (Hall, 1999: 438). The nature of working-class social capital is also different from that of other classes:

   while social clubs and trade unions dominate the
   associations to which working class members belong,
   those in the middle class develop affiliations with a much
   wider range of organisations.... The patterns of informal
   sociability of the working class are [also] more likely
   than those of the middle class to revolve around close
   contacts with kin and with a small set of friends ... (ibid.) (6)

From a historical-materialist perspective, a major factor underlying these class differences in the stock and nature of social resource in the form of social capital is the fact that the working class is excluded from, and relatively deprived of, the material resources (especially means of production) of society. A fundamental aspect of working-class life is also insecurity, not just exclusion from material resources. Because the working class has been separated from means of production, its access to means of subsistence, generally, depends on securing wage work which, however, is not guaranteed. One never knows how one is going to get through to the end of the week or the month. The material insecurity of the working class can undermine the conditions for reproduction of social-cultural norms of self-help and reciprocity, which presuppose a sense of security--an expectation that when one is in difficulty, one's relations with others will act as an insurance. Indeed, in a recent work, Putnam himself recognises this (although this recognition is not more than an afterthought in his overall framework, which treats social capital as an independent variable): as more and more workers are joining the contingent labour force (part-timers, people with insecure jobs, etc.), the conditions for connectedness at work are being undermined (Putnam, 2000: 88).

Further, given their separation from means of production and subsistence, working-class people are more likely to be in poverty than the propertied class. There is also much competition for jobs. Too much poverty and intra-class competition can adversely affect norms of reciprocity (Evans, 1997b; Brehm & Rahn, 1997: 1009); and there is also the issue of identity, which has been rightly emphasised in the postmodern/post-structuralist literature. Reproduction of the structure of social relations partly depends on agents' perceptions of themselves and of their 'others'--perceptions that underlie the formation of identities. Workers' crosscutting identities (ethnicity, gender, locality, etc.) can weaken their social capital in the form of their social networks (Callinicos, 1993: 50). White labourers may unite with white property owners in one area against minority workers in the same area. Place-bound, cross-class alliance is important: labourers in a city enter into alliance with capitalists in that city to attract inward investment that might hurt workers elsewhere (Fitzgerald, 1991; Harvey, 1982). These divisions can endure, as the capitalist class often uses them to control the working class (Wright, 1982: 539).

Now consider the macro social capital of the working class. I define this as norms of trust and reciprocity, and as networks and organisations, all located in the 'zone' of interaction between the working class and relatively autonomous pro-worker state officials. (It is important not to include the state as such as part of workers' social capital.) It refers to the 'synergy' (Evans, 1997b; pace Woolcock, 1998) between them. (7) In turn, 'synergy' refers to the provision of collective goods by the state in working-class communities, such as infrastructure and material assistance to the poorer workers. The synergy also refers to the state's embeddedness in these communities; i.e. to the ties of trust and reciprocity between state officials and working-class communities under their jurisdiction. Their participation in the daily lives of the working-class communities can enhance the latter's trust in officials. This also can make it possible for those communities to shame and censure officials when they are found to be working against workers' interests.

Macro-level social capital in the form of synergy between the state and the working class, like micro-level social capital, is both enabled and constrained by class processes. First of all, synergy is made possible by the existence of common grounds and common interests between the state (officials) and the working class. To the extent that the state represents capitalists' common interests (Miliband, 1977), the common grounds between the state and the working class partly reflect those that exist between the capitalist and working classes in particular places. Where these common interests also coincide with the relatively independent interests of the state (officials), the chances for cooperative action by the state and workers are maximised.

There are several possible common grounds or interests. First, both capitalists and workers, generally, have an interest in the weakening of pre-capitalist class structures and political relations (Rueschemeyer et al., 1992). The weakening of traditional structures of political domination also helps the state's penetration of society and its exercise of what Mann (1984) calls 'infrastructural power' (including that of rule-making). A local history of state-worker alliance against traditional elites can contribute to workers' macro social capital (at the local level). Second, an improvement in the material standard of living of the working class can be compatible with (and sometimes even might promote) capitalist economic growth. This will especially be the case where economic growth is based on the production of wage goods in the domestic market. The state also has an interest in growth because of its own dependence on that from which it receives its revenue.

Third, sometimes common ground between the working class and the state can be due to intra-capitalist class differences. For example, factions of the capitalist class can reap place-specific super-profits by paying below-subsistence wages to free labourers, and/or by using unfree labour (Brass, 1999). So state--worker synergy (through, for example, pro-worker legislation made possible by reformist officials) can make possible some increase in wages, but will leave the conditions for average, competitive profit rates for the capitalist class intact. The abolition of unfree labour in the same process will benefit labour, but will not be incompatible with the interests of the capitalist class as a whole, which generally presupposes free labour. The abolition of unfree labour, in so far as it allows the greater spatial mobility of the working class, might indeed benefit some capitalists, especially those located in areas of labour shortage.

These common grounds--and arguably there are others (see Wright, 2000)--exist structurally. They are to be translated into relations of cooperation between workers and the state through individual and collective agency. Following Poulantzas (1978: 148), I see the state as a socio-spatial site of contradictory class presences: a site in which both dominant and subordinate classes are present at different geographical scales, quite unequally and with vastly unequal amounts of influence. There are parts of the state, especially at the local scale, where pro-working-class elements or officials representing working-class interests are present. Indeed, as Poulantzas says, dominated classes are present in the state 'in the form of centres of opposition to the power of the dominant classes' (1978: 142). Such officials can take a reformist stance--'reformist' in that they may try to provide some temporary relief to workers within the constraints of capitalism/class. Whether their intention is to remain just reformist is beside the point here. The relations of trust and cooperation between reformist officials--including those who may have working-class backgrounds and ideology--and the working class itself can become a form of macro social capital of the latter. These officials can employ the autonomy of the state, using the state's resources at their disposal to try to empower working-class people, as they sometimes do for peasant/semi-proletarian communities (Fox, 1996; Bebbington & Perreault, 1999). The possibility that conflicts over wages and working conditions can be resolved (even if temporarily) in favour of the working class in specific places is enhanced by the presence of reformists within the state.

Indeed, Heller (1997) shows how social capital was produced in the Indian state of Kerala by both labour mobilisation and supportive state action. In particular, government action supported workers' mobilisation and offered institutional resources that allowed militancy to become compatible with accumulation. The implication of the synergy is that it made labour realise the need for discipline, as well as for reconciling redistribution with capitalist growth. Persistent political competitiveness led to a situation in which almost all parties, whether in or out of the government, were more or less committed to political mobilisation and the construction of encompassing organisations amongst working-class people.

But there are severe limits to the extent to which synergy between the state and workers can produce macro social capital. First, the state, partly because of its institutional separation from the capitalist economy (Offe, 1987), may not have the resources for the reformist officials to make available to the working class. Even if it does, it may be simply constrained from providing material benefits to the working class, given the influence of neoliberal ideology and policy. Without resources, norms of reciprocity between the state officials and workers will weaken. Second, even if the state does provide collective material and intangible goods, it may do so in a bureaucratic manner that can disrupt micro social capital in the form of workers' local initiatives and organisations (e.g. their self-help networks). This, in turn, can hurt the prospects of macro social capital formation. The bureaucratic distribution of benefits unequally between working-class members or areas can fragment the working class into individual citizen-clients competing for benefits from the paternalist state. Further, formal equality before the law and citizenship can, paradoxically, fragment the working class into individual citizen-voters, thus undermining class-based identity and solidarity. Those possibilities have been observed to have become realities in several countries, such as in Mexico and India (Fox, 1996; Das, 2000).

Third, the structure of the state itself makes synergistic interaction between the state and workers problematic. The state is not a neutral instrument equally accessible to all social forces. It has an in-built bias that makes it more open to dominant class influences. The state becomes a structurally selective terrain, which tends to insulate it from working-class influence on, and surveillance over it (Das, 1996: 50).This problem is compounded in the less developed nations by the fact that there, states become particularly repressive when they lack the resources with which to buy consent from the working class.

Finally--and this point applies to both the micro and macro social capital of the working class--when working-class social capital increases workers' material and political power beyond a level that dominant classes will tolerate, and if it becomes a political force in more than a few localities, the dominant classes will try to counter it by using their own social capital, including their ties with state officials. Beyond certain limits, the state and dominant classes will not tolerate cooperative action between reformist officials and workers, since the state's primary task is to reproduce the conditions of class exploitation and class domination. Neither will relations of cooperation within the working class, which can promote its unity vis-a-vis dominant classes, be likely to be supported by dominant classes. Working-class social capital is not necessarily the same as, but is a part of working-class collective action, in the sense that the former (and especially macro social capital) works within the limits of the class structure, whereas the latter can and must be aimed at transcending class relations. And there are practical as well as epistemological limitations of the concept of social capital. (8) If the function of social capital is social cohesion; if it is 'a glue that holds societies together' (Serageldin & Grootaert, 2000: 44), then working-class social capital as conceptualised here cannot be the social capital of social-capital researchers. Indeed, in the language of mainstream social-capital literature, it may be called 'negative social capital', or 'the dark side of social capital' (Putnam, 2000: 350-363). But within the limits of capitalism, a variable amount of working-class social capital can be tolerated by the capitalist class and the capitalist state.

6. Conclusion

Working-class social capital is about mutual relations of trust and cooperation within working-class communities--relations that can act as a social resource in promoting working-class solidarity, and therefore help the working class to contest the power of dominant classes. It is also about relations of trust and cooperation between workers and reformist officials. The social capital of the working class in both these senses can provide some material benefits to it.

In contrast to, but also complementing much of the class-struggle literature, which focuses mainly on the struggle between capital and labour, the approach in this paper allows one to look at opportunities for and constraints on: a) collective action and cooperation within the working class; and b) cooperation between the working class and reformist state officials. In contrast to approaches to social capital that generally under-theorise the class character of social capital, and overlook the literature on working-class self-activity, including self-help, this paper has attempted to show that social capital (like all socio-cultural practices) must be seen in its class context. This approach leaves open the possibility that workers can/do use social capital produced in a class-neutral context; that is, social capital as it is generally conceptualised in the literature; and this may give a semblance of similarity between working-class social capital and social capital per se (i.e. social capital as it is generally discussed in the literature). But my approach is distinctive from non-class approaches in recognising the severe limits to working-class social capital formation--the limits as defined by the class structure. Social capital scholars for whom social capital is a matter of networks/organisations that cut across class (and other) cleavages (Putnam, 1993: 175) do not have to worry about these limits. Social capital seen in a cross-class context is potentially very conservative. This form of social capital--cross-class social capital--is constituted by mutual relations of trust and connections between, say, the capitalist class and the working class, or between landlords and semi-proletarians (peasants), and is supposed to produce small benefits for the subordinate classes in specific places and for some time, while helping to reproduce the overall class structure at the national scale. On the other hand, social capital seen in a class context, as in working-class social capital, prioritises class conflict and has the potential to produce working-class solidarity vis-a-vis dominant classes.

I have argued that class is enormously constraining of social capital--a view often neglected in the literature on social capital, which treats social capital as an independent variable and a solution to social ills. Therefore it is absolutely problematic to try to construct a new theory of social capital of poverty, of health or crime (Woolcock & Narayan, 1999). There cannot be a social-capital theory of society (Das, 2004); but within a class theory of society, social capital can play some role. Given the constraints of class structure, generally, it will be a pretty minor role; although how minor that role will be depends on the specific issue at hand. It will also be place-specific, because the balance between class constraints and class opportunities--and indeed, the balance of power between classes--will vary geographically.

The approach outlined here is potentially open to some challenges from radical scholars. For example, a historical-materialist, whether she supports the social-capital idea or not, might argue that the state, given its class character, cannot help the working class much. However, underlying my approach to social capital is a specific approach to the state. In this approach, as far as workers' immediate interests, as opposed to their fundamental interests are concerned, state policies vary in their classness. ('Fundamental interests' are defined across modes of production [e.g. interests in capitalism or socialism]. 'Immediate interests' are defined within a mode of production [Wright, 1982: 541].) Thus, some policies are more pro- or anti-working-class than others; and within the limits of capitalist property relations and the logic of accumulation, states' actions are generally a product of, and reflect working-class power and struggle. Within the limits of class structure, therefore, certain synergistic relations between working-class people and reformist state officials are possible (especially at the local level). Further, capital depends on the state's protecting exploitative property rights through a combination of coercion and consent. This fact might be sufficient to make the capitalist class tolerate a certain amount of working-class (macro) social capital.

Second, one could argue that my approach is reformist. My response would be that, if anything, it is certainly much less reformist than the non-class approach to social capital, which believes in the trickle-down theory of social capital, ignoring the specificity of working-class social capital. Besides, the approach is not automatically inconsistent with working-class radicalism. There is nothing necessarily reformist about stressing relations of trust and reciprocity within the working class; that is, among individual workers and among groups of workers. (In a sense, we don't 'see' the working class as such; we see groups of workers and individual members of the working class, in given places, who act.) A class approach to social capital and the attendant emphasis on working-class social capital have indeed some potential for radical politics. To the extent that working-class people perceive that possibilities for reform within capitalism exist unexploited, that perception is a great danger to the working class's struggle over its fundamental interests. This is arguably the 'political' counterpart to Marx's 'economic' claim that 'No social order ever perishes before all the productive forces for which there is room in it have developed' (Marx, 1977: 390). An important underlying aim of my approach is to make it possible for one to see the severe class-structural limits to working-class synergy with the state, and to working-class members' practicing of self-help. Another implication is that the construction of social capital itself is a process of political awareness of the working class in the sense that Marx and Luxemburg recognised: through mutual aid and various working-class associations, including those which are 'purely' for sociability, this political awareness is raised. Arguably, relations of trust and cooperation among working-class people, especially if these are informed by class consciousness, provide a foundation for political solidarity. Given the global neoliberal attack on the working class, it is very important that all avenues of promoting working-class political awareness and organisation be employed in particular places and everywhere. Working-class social capital is only one of them. Capitalists use their social capital, i.e. their informal connections with one another and with state officials, in their mutual struggle and in their struggle against the working class. So, the working class must also produce and use its own social capital. (9)

As I have said, much of the social-capital literature reflects the current worldwide neoliberal agenda. (10) For one thing, its stress on mutual assistance and self-help, which NGOs (with support from international agencies such as the World Bank) tend to fetishise, reflects and reinforces an important neoliberal idea. This is that the burden of reproduction of lower-income families whose jobs are lost through capitalist restructuring should be borne by those families themselves, through such processes as self-help. It should not be borne through a reduction in the capitalist class's share of the total social product that workers produce. In particular, it should not be borne by the state, whose welfare role must be curtailed (Wills, 2000: 646). The argument in this paper that there are severe class limits to the reproduction of norms of reciprocity between working-class people, and to their practices of mutual aid, actually flies in the face of the neoliberal idea that working-class people under the onslaught of a neoliberal attack can 'bootstrap' themselves up through mutual aid--an idea which, as Petras (1997) says, is very dear to many NGOS and post-Marxists. Further, social-capital research on how non-market processes such as mutual trust and customs grease the wheels of the market by reducing transaction costs (Serageldin & Grootaert, 2000: 48) fits in well with, and generally supports the neoliberal agenda of making markets more efficient--a euphemism for the augmentation of capitalist profits. By destroying working-class lives and places through such processes as job loss and the withdrawal of welfare provisions by the state, neoliberalism has created a potential for class confrontation (Zeilig, 2000; Petras, 1997). Cross-class social capital as a glue--indeed, Putnam describes it as a 'superglue' (Putnam, 2000: 23)--can be very helpful in avoiding this situation, potentially making social-capital believers, including their supporters in the World Bank, great social engineers. But a class approach to social capital has a different agenda in so far as it draws attention to working-class norms of reciprocity and its associations, and to the potential of working-class social capital as 'an oppositional social capital' (Body-Gendrot & Gittell, 2003: xiii; also see Warren et al., 2001: 6) to promote its political solidarity vis-a-vis capitalists and the state. Working-class social capital as conceptualised here is not a sociological superglue that holds conflicting elements of society together--although to the extent that it is a glue, it is one that can, to a degree, promote unity among working-class people, thus furthering their ability to confront the capitalist class and the state. And to the extent that it is a grease, it has some potential to grease the wheels, not of the capitalist market but of class conflict.

Aknowledgement

I wish to thank the four referees and the Capital & Class Editorial Board, and especially Dr Andreas Bieler, for their very helpful comments on an earlier draft between 2001 and 2003 of this paper. Although it contains very little empirical material on social capital, the paper developed partly through my critical reflections on the findings from my empirical research on the social capital of wage labourers, which was funded by the Economic and Social Research Council (Grant number R000223723) during the period 2001-2003.

References

Barr, A. (2000) 'Social capital and technical information flows in the Ghanaian manufacturing sector' in Oxford Economic Papers, vol. 52, no. 3, pp. 539-559.

Bebbington, A. & T. Perreault (1999) 'Social capital, development, and access to resources in Highland Ecuador' in Economic Geography, vol. 75, no. 4, pp. 395-418.

Bodeman, Y. & W. Spohn (1986) 'The organicity of classes and the naked proletarian: Towards a new formulation of the class conception' in The Insurgent Sociologist, vol. 13, no. 3.

Body-Gendrot, S. & M. Gittell (2003) 'Empowering citizenship to social capital' in S. Body-Gendrot & M. Gittell (eds.) Social Capital and Social Citizenship (Lexington Books).

Bourdieu, P. (1986) 'The forms of capital' in J. Richardson (ed.) Handbook of Theory and Research for the Sociology of Education (Greenwood).

Brass, T. (1999) Towards a Comparative Political Economy of Unfree Labour (Frank Cass).

Brautigam, D. (1997) 'Substituting for the state: Institutions and industrial development in eastern Nigeria' in World Development, vol. 25, no. 7, pp. 1063-1080.

Brehm, J. & W. Rahn (1997) 'Individual-level evidence for causes and consequences of social capital' in American Journal of Political Science, vol. 41, no. 3, pp. 999-1023.

Callinicos, A. (1993) Race and Class (Bookmarks).

-- (1988) Making History: Agency, Structure and Change in Social Theory (Cornell University Press).

Camfield, D. (2004) 'Re-orienting class analysis: Working classes as historical formations' in Science and Society, vol. 68, no. 4, pp. 421-446.

Chhibber, A. (2000) 'Social capital, the state, and development outcomes' in P. Dasgupta & I. Serageldin (eds.) Social Capital: A Multifaceted Perspective (World Bank).

Dahl, R. (1956) A Preface to a Democratic Theory (University of Chicago Press).

Das, R. J. (2004) 'Social capital and poverty of wage labourers: Problems with the social capital theory' in Transactions of the Institute of British Geographers, vol. 29, no. 1, pp. 27-45.

-- (2000) 'State-society relations: The case of an antipoverty policy' in Environment and Planning C: Government and Policy, vol. 18, no. 6, pp. 631-650.

-- (1996) 'State theories: A critical analysis' in Science and Society vol. 60, no. 1, pp. 27-57.

Duncan, C. (2001) 'Social capital in America's poor rural communities' in Social Capital and Poor Communities (Russell Sage Foundation).

Evans, P. (1997a) 'Introduction: Development strategies across the public--private divide' in P. Evans (ed.) State--Society Synergy (University of California Press).

-- (1997b) 'Government action, social capital, and development' in P. Evans (ed.) State--Society Synergy (University of California Press).

Fine, B. (2001) Social Capital vs. Social Theory (Routledge).

-- (2000) 'Social capital: A critique', from an email discussion list about the 2000-2001 World Development Report, at <[e-mail]>.

-- (1998) 'The developmental state is dead: Long live social capital?' in Development and Change, vol. 30, no. 1, pp. 1-19.

Fitzgerald, J. (1991) 'Class as community: The new dynamics of social change' in Environment and Planning D, vol. 9, no. 1, pp. 117-128.

Foley, M. & B. Edwards (1996) 'The paradox of civil society' in Journal of Democracy, vol. 7, no. 3, pp. 38-52.

Fox, J. (1996) 'How does civil society thicken? The political construction of social capital in rural Mexico' in World Development, vol. 24, no. 6, pp. 1089-1103.

Granovetter, M. (1985) 'Economic action and social structure: The problem of embeddedness' in American Journal of Sociology, vol. 91, no. 3, pp. 481-510.

Greene, A., P. Ackers & J. Black (2001) 'Lost narratives? From paternalism to team-working in a lock-manufacturing firm' in Economic and Industrial Democracy, vol. 22, no. 2, pp. 211-237.

Grootaert, C., G. Oh & A. Swamy (2002) 'Social capital, household welfare and poverty in Burkina Faso' in Journal of African Economies, vol. 11, no. 1, pp. 4-38.

Grootaert, C. & D. Narayan (2004) 'Local institutions, poverty and household welfare in Bolivia' in World Development, vol. 32, no. 7, pp. 1179-1198.

Hall, P. (1999) 'Social capital in Britain' in British Journal of Political Science, vol. 29, no. 3, pp. 417-461.

Harvey, D. (1982) Limits to Capital (Chicago University Press).

Heller, P. (1997) 'Social capital as a product of class mobilization and state intervention: Industrial workers in Kerala, India' in P. Evans (ed.) State--Society Synergy (University of California Press).

Herod, A. (1997) 'From a geography of labour to a labour geography: Labour's spatial fix and the geography of capitalism' in Antipode, vol. 29, no. 1.

Kennedy, B., I. Kawachi & E. Brainerd (1998) 'The role of social capital in the Russian mortality crisis' in World Development, vol. 26, no. 11, pp. 2029-2043.

Knack, S. & P. Keefer (1997) 'Does social capital have an economic payoff? A cross-country investigation' in Quarterly Journal of Economics, vol. 112, no. 4, pp. 1251-1288.

Li, Y., A. Pickles & M. Savage (2005) 'Social capital and social trust in Britain' in European Sociological Review, vol. 21, no. 2, pp. 109-123.

Levi, M. (1996) 'Social and unsocial capital: A review essay of Robert Putnam's Making Democracy Work' in Politics and Society, vol. 24, no. 1, pp. 45-55.

Marx, K. (1975 [1847]) The Poverty of Philosophy (Progress Publishers).

-- (1967 [1852]) The Eighteenth Brumaire of Louis Napoleon (Progress Publishers).

Morris, M. (1998) 'Social capital and poverty in India', Institute of Development Studies working paper, University of Sussex, no. 61, p. 25.

Morris, L. & S. Irwin (1992) 'Unemployment and informal support: Dependency, exclusion, or participation' in Work, Employment and Society, vol. 6, no. 2, pp. 185-207.

Muntaner, C., J. Lynch, M. Hillemeier, J. Lee, R. David, J. Benach & C. Borrell (2002) 'Economic inequality, working-class power, social capital, and cause-specific mortality in wealthy countries' in International Journal of Health Services, vol. 32, no. 4, pp. 629-656.

Ostrom, E. (2000) 'Social capital: A fad or a fundamental concept' in P. Dasgupta & I. Serageldin (eds.) Social Capital: A Multifaceted Perspective (World Bank).

Petras, J. (1997) 'Imperialism and NGOS in Latin America' in Monthly Review, vol. 49, no. 7, pp. 10-27.

Portes, A. & J. Sensenbrenner (1993) 'Embeddedness and immigration: Notes on the social determinants of economic action' in American Journal of Sociology, vol. 98, no. 6, pp. 1320-1350.

Poulantzas, N. (1978) State, Power, Socialism (New Left Books).

Ruschemeyer, D., E. Stephens & J. Stephens (1992) Capitalist Development and Democracy (Polity).

Serageldin, I. & C. Grootaert (2000) 'Defining social capital: An integrating view' in P. Dasgupta & I. Serageldin (eds.) Social Capital: A Multifaceted Perspective (World Bank).

Short, P. (1996) 'Kinship, reciprocity and vulnerability: Social relations in the informal economy' in Australian Journal of Social Issues, vol. 31, no. 2, pp. 127-145.

Truman, D. (1959) The Governmental Process (Knopf).

Uphoff, N. & C. Wijayarama (2000) 'Demonstrated benefits from social capital: The productivity of farmer organizations in Gal Oya, Sri Lanka' in World Development, vol. 28, no. 11, pp. 1875-1890.

Warren, M., J. Thompson & S. Saegert (2001) 'The role of social capital in combating poverty' in S. Saegart, J. Thompson & M. Warren (eds.) Social Capital and Poor Communities (Russell Sage).

Wood, E. (1998 [1986]) The Retreat from Class (Verso).

Woolcock, M. (1998) 'Social capital and economic development: Toward a theoretical synthesis and policy framework' in Theory and Society, vol. 27, no. 2, pp. 151-208.

World Bank (2001) World Development Report: Attacking Poverty (Oxford University Press).

Wright, E. (2000) 'Workers' power, capitalist-class interests, and class compromise' in American Journal of Sociology, no. 105.

-- (1995) 'The class analysis of poverty' in International Journal of Health Services, vol. 25, no. 1, pp. 85-100.

-- (1989) 'Rethinking, once again, the concept of class structure' in E. Wright (ed.) The Debate on Classes (Verso) pp. 269-348.

-- (1982) 'Race, class and income inequality' in A. Giddens & D. Held (eds.) Classes, Power and Conflict (University of California Press).

Zeilig, L. (2002) Class Struggle and Resistance in Africa (New Clarion).

Notes

(1.) Social-capital theory is just a variety of pluralism (see Dahl, 1956; Truman, 1959), since pluralism is about cross-cutting cleavages and ignores class formation. I am grateful to one referee for pointing this out to me.

(2.) Social capital can appear to be social capital of the poor only when social capital's class context is stripped away. It appears progressive when the 'power, securities, and opportunities of the wealthy [that] remain sacrosanct' are abstracted from (Fine, 2000).

(3.) I am aware of the problematic nature of 'social' in social capital. 'Social' in social capital implies that 'capital' is not necessarily social, when in fact it is. I am using the word 'capital' in 'social capital' in the sense of a resource, rather than as 'capital' in the Marxist sense, as noted above. One might argue that I am suggesting that there is a common pool of social capital, of which the working class must have a share like the capitalist class. I am not saying this. Relations of connections and trust and their causal powers are specific to classes (and genders and ethnic groups). Relations of connection among people from the capitalist class constitute their social capital, and will have little use for the working class, although they are structured by the same class relations that structure relations of connections between working-class people. So it is not a matter of a society-wide stock of social capital that different classes fight over. If anything, each class has its own (class-specific) stock of social capital, which develops often in antagonism with the social capital of another class.

(4.) As my own empirical research in India shows (Das, 2004: 32-33; 39-40), the informal borrowing of food can act as a cushion for poor working-class households against employment/income insecurity, which is a serious problem under the neoliberal regime of employment shrinkage and the withdrawal of welfare provisions. No one knows whether they will work on a given day or not, and whether they will thus be able to buy some food. But the fact that they can borrow some food from their neighbours--and this borrowing is facilitated by working-class social capital (norms of reciprocity)--means that they may escape from starvation. But there are severe limits to social capital as a cushion against insecurity, as I discuss below.

(5.) Thus workers' micro social capital is produced by ties between individual workers and groups of workers. Some of these ties are ties of kinship, family, location, and so on. Social capital produced by these ties between workers can contribute to their unity, and provide temporary forms of material support. More importantly, these ties come from their 'emergent sentiment of we-ness' or 'bounded solidarity'--class consciousness--which is, in turn, reinforced by these ties and associational life. Further, workers' social capital is not the same as class consciousness, but it can be helped by it. Workers' social capital is also not the same thing as their political solidarity, but it can be based on solidarity and reinforce it. Neither is workers' social capital the same as class conflict--but it can help it to materialise.

(6.) Although classes in Hall's study are defined following the Weberian methodology, the results are broadly indicative of the class differences in social capital that this paper discusses.

(7.) I am deliberatively excluding from macro social capital the relations of trust and community between workers and managers/capitalists, or what Greene et al. call 'psychological contract'. The features of psychological contract include: workers identifying themselves with companies; face-to-face communication between managers and workers on the shop floor and in informal social occasions in the community, outside of work; joint decision-making and consultative committees; workers' expectation of managers to play a paternalist role; and a sense of community at the workplace (Greene et al., 2001: 229-231). A mainstream social-capital writer who believed in social capital as a glue, in contrast, might include these features under the heading 'social capital'.

(8.) To the extent that social-capital writers discuss limitations of social capital as a practice (not as a concept), as in the discussion of the dark side of social capital (e.g. mafia social capital), the limitations are contingently related to the social-capital practice. In a class context, the limitations are a necessary aspect of the practice. It is also the case that some social-capital writers such as Putnam occasionally recognize some of the limitations I discuss, but then proceed as if these did not exist.

(9.) Working-class unity for political action is easier when workers themselves are connected through personal ties, especially if those ties are informed by class consciousness. Class-based 'bounded solidarity' can produce these personal relations (social capital); but these personal relations of trust and cooperation will not necessarily produce solidarity vis-a-vis dominant classes.

(10.) However, it is to be noted that not all social-capital enthusiasts are necessarily neoliberals, and that anarchists and populists, whose views may be anti-neoliberal, may be sympathetic to social capital.