Is a focus on social capital useful in considering food security interventions? Insights from KwaZulu-Natal

Abstract

Social capital is an important collective resource people draw on in pursuit of well-being. This article explores the nexus between household social capital and food security in a small community in KwaZulu-Natal. The case study suggests some social capital-related failures are linked to food insecurity in the community, including a breakdown in two-parent families, divergences between religious groups, ambiguous leadership characterised by conflict, and changes in cultural norms. The highly variable and household-specific nature of social capital's role in food security makes it difficult to extrapolate lessons for targeting social capital in food security interventions beyond the case-study community. However, the findings point to the value of including proxies for social capital in vulnerability indices and food-insecurity mapping systems, and more broadly to the importance of understanding context-specific interactions between resources or ‘capitals’, institutional issues, and the human relationships and power dynamics that shape food insecurity and the outcomes of interventions to address it.

Keywords:

• Food security
• social capital
• vulnerability
• KwaZulu-Natal

1. INTRODUCTION

A substantial amount of research evidence suggests that people's stores of social capital play a vital role in defining not only their resilience to livelihood shocks of all kinds, but also their sustainable development toward economic and social well-being (Narayan & Pritchett, 1997; Fafchamps & Minten, 1998; Maluccio et al., 1999; Kliksberg, 2000; Woolcock, 2000; Isham, 2001). Food security status is a significant dimension of human well-being. Southern Africa experiences the highest prevalence of food insecurity in the world, as evidenced by the prevalence of malnutrition, persistent famine and food crises, and declining per-capita food production (Food and Agriculture Organisation of the United Nations, 1993; Rosegrant et al., 2001; Meissner, 2002; Morris, 2002). Even in southern Africa's wealthiest country, South Africa, food insecurity is such that it has been acknowledged as a national priority for government policy development and public spending since 1994 (Department of Agriculture & Land Affairs, 1997).

The development of effective food security interventions in South and southern Africa has, however, arguably been hampered by entrenched ‘disaster management’ rather than long-term development approaches, together with a preoccupation with food production deficits as the primary determinant of food insecurity (Holloway, 2003). Most commentators now agree that the real challenge in southern Africa is to address the food insecurity of ‘creeping’ vulnerability, which is rooted in structural socio-economic and political factors that lead to ongoing livelihood failures and food insecurity, and which may lower resilience to environmental hazards (Devereux, 2001; De Waal & Tumushabe, 2003; Holloway, 2003; Misselhorn, 2005). The purpose of much work in both the academic and development spheres has been to understand these underlying conditions better and to suggest how they might be addressed, particularly within the broader framework of vulnerability studies.

The concept of social capital represents important elements embedded in the underlying and broader socio-economic context of food insecurity. Community social characteristics that fall under the umbrella of social capital include relations of trust, reciprocity and exchange, common rules, norms and sanctions, social connectedness, and social networks and groups (Adger, 2003; Pretty, 2003). These characteristics have been argued to function in critical ways to mitigate aspects of vulnerability and to build livelihoods in communities (Narayan & Pritchett, 1997; Fafchamps & Minten, 1998; Maluccio et al., 1999; Woolcock & Narayan, 1999; Kliksberg, 2000; Scoones, 2000; Isham, 2001). Recent research in southern Africa has suggested that failures in social capital resources contribute substantially to failures in household-level food security (De Waal & Tumushabe, 2003; Drinkwater, 2005; Misselhorn, 2005).

In the light of the above, can – or should – decision-makers, who are currently investing tremendous resources in mitigating food insecurity, take social capital into consideration in the development of interventions? This is the central question that is explored in this article through an examination of the results of a case study in a small community in the province of KwaZulu-Natal.

2. SOCIAL CAPITAL

2.1 Perspectives on social capital

Social capital is in many respects a controversial idea and this article does not aim to theorise it extensively. It is necessary, however, to briefly review some perspectives and criticisms.

Three general perspectives on social capital are summarised here: the communitarian view, the networks view, and the institutional view (Grootaert, 1998; Woolcock & Narayan, 1999). The communitarian view narrowly equates social capital with the density of local-level organisations in any given community. The networks view emphasises that social capital may have both positive and negative outcomes, and the ‘tension’ between these is one of this perspective's defining properties (Woolcock & Narayan, 1999:7), a tension that is characterised by the balance between ‘bonding’ and ‘bridging’ ties. ‘Bonding’ social capital usually denotes close relationships such as those between family and friends, while ‘bridging’ social capital denotes more distant relationships in which people are ‘weakly’ tied by very context-specific interaction (e.g. Gittel & Vidal, 1998).

The third perspective, the institutional view, extends the networks view to incorporate the role of a community's social and political environment, recognising that a community's links to formal (external) institutions, particularly state institutions, determine the capacity of its internal institutions to advance collective community well-being. A distinction is drawn between ‘horizontal’ and ‘vertical’ (or ‘hierarchical’) relationships (e.g. Putnam, 1993; Moser & Holland, 1998), suggesting that a key element of successful community development is the vertical social networks beyond the community scale that ultimately determine the political power of communities and their leverage in state–civil society relationships (e.g. Woolcock & Narayan, 1999). Relations of power between individuals, households, communities and organisations are recognised as endemic to institutional structures and processes (Ife, 1995), and are thus understood here to be inherent in the idea of social capital.

While all institutions at all scales fall under the general umbrella of ‘social capital’, the theory of social capital extends to include the interactions and communications between people that feed communities' capacity to build and invest stores of social capital (see, e.g., Falk & Harrison, 1998; Gittel & Vidal, 1998; Daubon & Saunders, 2000; Narayan & Cassidy, 2001).

2.2 Criticisms of the concept of social capital

The concept of social capital has been criticised by many scholars. The first area of concern is the problematic nature of the concept itself. Part of the attractiveness of the idea is that it encapsulates a suite of intuitively important social resources that are indisputably significant in shaping livelihoods (Portes & Landolt, 1996; Narayan & Pritchett, 1997; Fafchamps & Minten, 1998; Maluccio et al., 1999; Woolcock & Narayan, 1999; Kliksberg, 2000; Isham, 2001). Recently, a more critical analysis of social capital has been called for, so that opportunities for more directed research towards increased social adaptive capacity might be pursued (Pelling & High, 2005).

A second area of criticism comes from research that finds communities that have extensive group membership and solidarity may still fail to overcome crippling structural constraints to development (Portes & Landolt, 1996; Flora, 1998; Woolcock & Narayan, 1999). In other research, social capital has been found to have the capacity to encourage ‘pathological’ community responses. For example, strong group exclusivity or homogeneity may lead to ethnic or racial divides, or corrupt or violent community elements (Rubio, 1997; Mohan & Mohan, 2002). Different forms of social capital, and the different ways these forms are used, have different outcomes, as research in the US indicates (Schorr, 1992; Putnam, 2000; Putnam & Feldstein, 2003). Related to this, a final area of concern is the question of whether, and how, policy interventions can foster social capital to the benefit of society (American Civic Forum, 1994; Starr, 1994; Sirianni & Friedland, 1995; Pelling & High, 2005).

The rise of social capital as a development concept has been fraught with concerns and controversy over its usefulness and meaning. This article highlights and discusses a number of more specific difficulties that were experienced in attempting to probe the relationship between household social capital and food security in the case-study community.

3. KWAZULU-NATAL CASE STUDY

3.1 Rationale for case-study locality

The high prevalence of HIV/AIDS in KwaZulu-Natal, South Africa (16.5 per cent of the population over 2 years) and the multiple links between HIV/AIDS and food insecurity (Morris, 2002; Piot et al., 2002; De Waal & Tumushabe, 2003; Mano et al., 2003; US Agency for International Development, 2003; Human Sciences Research Council [HSRC], 2005) suggest that in coming years food insecurity will continue to deepen in the province (Cross, 2002; HSRC, 2002; Drimie, 2003).

KwaZulu-Natal's poverty characteristics indicate a particularly high risk of food insecurity. The province has the second highest concentration of poor rural residents after the Eastern Cape (HSRC, 2004). Almost 87 per cent of the ‘chronically poor’ in KwaZulu Natal live in rural areas (Aliber, 2003), and the rural poor in South Africa are seen as experiencing food insecurity more acutely than those in urban areas (Statistics South Africa & United Nations Development Programme, 2003). The extent of those affected by poverty in the province also increased from 26.8 per cent in 1993 to 42.5 per cent in 1998 (according to the KwaZulu-Natal Income Dynamics Study) and, unlike in other provinces, the severity of poverty rose over the same period (Carter & May, 2001; Adato et al., 2004). KwaZulu-Natal also has by far the widest aggregate poverty gap of all provinces in South Africa (HSRC, 2004).

The province's HIV/AIDS prevalence and poverty characteristics suggested the selection of a community in this province as the focus of this case study. The exact identity of the community selected for the case study is withheld in this article for ethical reasons, but it lies in a valley some kilometres from Durban, and just a few kilometres by road from the urban centre of Pinetown. Formerly part of the KwaZulu-Natal ‘homeland’ under apartheid, it now falls under the Ethekwini Metropolitan area, thus being legally classified as urban after the Municipal Demarcation Act, No. 27 of 1998 (Lehohla, 1999). Estimates by community members put the total number of households in the community at between 300 and 500. Although classified as urban, in reality the community is semi-rural in character, with about 80 per cent of households engaging in some form of agriculture. The case-study results reveal very low rates of formal employment here (about 46 per cent of adults) and little engagement in informal income-generating activities (only three adults out of 157 living in the surveyed households). Nearly one-fifth of the households surveyed here rely solely on one or more government grants for income. Low levels of education are prevalent among household members, with about 21 per cent of adults in the survey reportedly having no formal education.

3.2 Research methods

Multiple methods were used to give depth and contextual breadth to the findings (Ragin, 2000; Marsland et al., 2001; Schutze et al., 2004). These methods included a questionnaire survey, semi-structured interviews undertaken at the outset of the field work followed by in-depth interviews with households selected on the basis of the survey results, a focus group with health workers and home-based carers, and the recording of anthropometric data (heights and ages of children).

A simplified multistage sampling technique was used to select the 50 households participating in the questionnaire survey (De Vaus, 1991; Carletto, 1999; Lehtonen & Pahkinen, 2004). The community was divided into five geographical areas estimated to have similar numbers of households, and a sample of 10 households was drawn randomly from each. However, since participation was voluntary, this precluded a strict adherence to administering the questionnaire to pre-selected households.

The heights and ages of 60 children at the two community crèches were recorded to determine their height for age. Being under-height-for-age is accepted as a possible indicator of undernourishment, which in turn may be a symptom, or indirect indicator, of food insecurity (Scherr, 2003).

Drawing on the theoretical constructs of the dimensions of social capital, and on the Sustainable Livelihoods framework (Scoones, 1998), a research framework was constructed in accordance with the research objectives. Four focus variables were included: food security, social capital, financial capital and human capital. The focus variable of food security was indirectly measured in the survey through the indicator of dietary diversity, which is among the most widely used and cost-effective of methods. The dietary diversity question required respondents to record the number of unique foods consumed over the previous 14 days, which yielded a dietary diversity score for each participating household (Krebs-Smith et al., 1987; Drewnowski et al., 1997; Hatloy et al., 1998; Hoddinott, 1999; Hoddinott & Yohannes, 2002; Ruel, 2002; Webb & Lapping, 2002; Holloway, 2003; Hendriks, 2005).

The survey covered seven broad aspects of the focus variable of social capital: (1) groups and networks, (2) trust and solidarity, (3) collective action and cooperation, (4) information and communication, (5) social cohesion and sociability, (6) conflict and violence, and (7) empowerment and political action. Particular emphasis was placed on trust and membership in groups since these have been better researched than other measures of social capital (Falk & Harrison, 1998; Narayan & Cassidy, 2001; Grootaert et al., 2004).

The financial capital focus variable was represented in the survey through a question probing employment levels in each household, and human capital was represented by reported household adult education levels.

Sixteen analysis indices, including the dietary diversity score, were collated from the questionnaire to represent the four focus variables, or aspects of these (Table 1). The relationship between food security and each of these indices was explored through a number of statistical techniques, including Pearson's product moment correlation coefficient, Kendall's tau, analysis of variance, regression analysis, and a multivariate cluster analysis (Kline, 1994; Sheskin, 2000; Guidici, 2003).

Table 1: 16 analysis variables synthesised from survey results and the focus variables they represent

Focus variables represented	16 analysis variables
Food security	Dietary diversity score
	Number of meals
Social capital	Groups and networks
	Trust and solidarity
	Collective action and cooperation
	Social cohesion
	Sociability
	Conflict and violence
	Empowerment and political action
Household livelihood strategies	Gender of household head
	Family size
	Crop volume
	Crop diversity
Financial capital	Ratio employed to family size
	Ratio grants to family size
Human capital	Adult education index
Note: The focus variable listed on the left is considered the primary variable represented by the analysis variable on the right. Some of the 16 variables represent more than one focus variable; for example, crop volume and crop diversity pertain to both food security and livelihood strategies.

4. SOCIAL CAPITAL OBSTACLES AND OPPORTUNITIES

4.1 The variable dynamics of household capitals

The case-study analysis suggests that livelihoods in the community are shaped by household-specific capitals, in which food security and other livelihood characteristics are determined by a unique combination of resources that are used differently, and that also interact with one another differently, in the context of different households.

The statistical analyses of the survey data revealed no distinct or consistent patterns among households between access to any form of capital, or to any combination of capitals (as represented in the 16 analysis indices synthesised from the raw data), and their levels of food security (Table 2). An overall perspective offered by a regression analysis similarly showed that only about 31 per cent of all the variance in food security was explained by all 16 predictor variables working together, and none of the variables individually made a significant contribution to the variation in food security at the 5 per cent level (Table 3). Further, only one cluster for the continuous variables was found in a two-step cluster analysis, suggesting that respondents in the survey were either generally similar in the way that they responded to questions or that they were too dissimilar to display patterns of similar response.

Table 2: Coping strategies used by households in response to income shocks

Coping strategy	Number of citations
Reduce amount of food bought or buy cheaper food	22
Borrow money	16
Get help from others	11
Sell assets or use savings	2
Take children out of school	2
Use insurance	0
Note: Some households experience more than one shock, and thus the same coping strategy may have been cited as being adopted more than once within the same household during the previous 3 years.

Table 3: Relationships between food security and 16 other analysis variables

Relationship between food security and …	Pearson's r	Approximate significance	Kendall's tau	Approximate significance
Ratio employed to family size	0.17	0.21	0.20*	0.04
Ratio grants to family size	0.12	0.41	0.12	0.31
Adult education index	0.15	0.29	0.12	0.28
Groups and networks	–0.04	0.66	0.00	0.98
Trust and solidarity	–0.04	0.77	–0.02	0.78
Collective action and cooperation	–0.01	0.90	–0.06	0.54
Social cohesion	0.24	0.09	0.22*	0.03
Sociability	–0.03	0.13	0.03	0.76
Conflict and violence	0.14	0.13	0.18	0.14
Empowerment and political action	0.00	0.99	0.03	0.79
Family size	0.18	0.22	0.06	0.60
Crop volume	0.13	0.52	0.03	0.82
Crop diversity	0.16	0.29	0.04	0.73
Number of meals^a	–0.33*	0.02	–0.28*	0.00
Gender of household head	0.18	0.21	0.12	0.31
^aThe weak but statistically significant relationship according to Kendall's tau between the ‘number of meals eaten per day’ in a household and food security (–0.28) suggests that families with lower levels of food security eat more frequently. In the analysis of variance analysis, the ‘number of meals eaten per day’ was found to have a variance that was not homogeneous with that of food security, and thus similarly had to be excluded from the analysis. *Statistically significant at the 5 per cent level, as indicated by the p values in the last column being below 0.05.

The inconsistent and uncertain impact of forms of social capital on food security that the statistical analyses suggested was further interpreted in the light of the in-depth household interviews. The statistical analyses and qualitative explorations suggest that social capital has a variable or inconsistent effect on food security, the production of financial capital, knowledge (human capital) and well-being. Capitals, or resources, in the case-study community clearly interact in complex ways to shape livelihoods and food security. Forms of social capital can either build or diminish other forms of capital, depending on household dynamics and a household's relationships within the community (Table 4). This finding of intricate interactions between capitals agrees with a ‘systems’ view (Berkes & Folke, 1991) of human and environmental interactions.

Table 4: Summary of results of regression analysis on all variables as predictors of food security

Variable	Standardised β coefficient	t value	Approximate significance	5% confidence interval for β
				Lower bound	Upper bound
Ratio employed to family size	0.13	0.70	0.50	–5.721	11.72
Ratio grants to family size	0.31	1.84	0.08	–0.51	10.30
Adult education index	0.09	0.54	0.59	–0.79	1.37
Groups and networks	–0.00	–0.01	0.99	–0.44	0.44
Trust and solidarity	–0.14	–0.81	0.43	–0.84	0.36
Collective action and cooperation	–0.15	–0.78	0.44	–0.97	0.43
Social cohesion	0.20	1.18	0.24	–0.64	2.44
Sociability	–0.12	–0.77	0.45	–0.67	0.30
Conflict and violence	0.15	0.89	0.38	–1.22	3.12
Empowerment and political action	0.14	0.82	0.42	–0.81	1.92
Family size	0.33	1.80	0.08	–0.09	1.45
Crop volume	0.01	0.03	0.98	–0.65	0.67
Crop diversity	0.09	0.23	0.82	–2.11	2.66
Number of meals	–0.25	–1.50	0.14	–5.48	0.82
Note: β value is standardised to enable comparison between variables as to their contribution to the variation of food security. For example, the ratio of employed to family size has a β value of 0.13, which indicates that for every unit of change in this index, there is a 0.13 proportional change in food security. None of the variables made a significant contribution (at the 5 per cent level), however, to variation in food security as shown by the p values all lying above 0.05.

The complexity of household capital dynamics is illustrated in the cases of those households with the two highest and two lowest dietary diversity scores (as proxy indicators of food security) (Table 5). Despite there typically being a strong linear relationship between food insecurity and poverty (World Health Organisation, 2002), this was not evident in the statistical analyses of the survey data. Analysis of the interviews with these households suggests that while financial capital is undeniably an important determinant of food security in all four households, the means of generating financial capital and the way it is used to pursue well-being and to generate more capital vary considerably. Stokvel groups, for example, significantly shape the ability of these households to access sources of financial capital. Stokvels are common in southern African communities. These groups usually provide some kind of rotating assistance. For example, in a savings stokvel, members contribute a small fixed sum of money, usually monthly, which is pooled and allocated as a lump sum to each group member in turn. Stokvels were a dominant community-level institution across the survey, with 68 per cent of households belonging to savings stokvels and 54 per cent to lending stokvels. Stokvel membership is clearly a valuable resource in Households One, Two and Three (Table 5). However, the ability of Household Four to participate in a stokvel was inhibited because of the household head's inability to contribute during a period of financial stress. This gave rise to significant conflict with other stokvel members and ultimately led to her exclusion from the group.

Table 5: Food security characteristics, capital resource availability and livelihood strategies of households with the two highest and two lowest dietary diversity scores

	Household One	Household Two	Household Three	Household Four
Food security characteristics	Highest food security index	Second-highest food security	Second-lowest food security index	Lowest food security index
	Eat well by own assessment	Eat well by own assessment	Do not eat good quality or sufficient quantity by own assessment	Do not eat good quality or sufficient quantity by own assessment
	Occasionally have to reduce amount of food due to shortages	Occasionally have to reduce the amount of food due to shortages	Occasionally reduce food due to shortages	When food is very short adults go without
	Knows three of the five food groups	Do not know the food groups	Do not know the food groups
Social capital	Two-parent family, two children	Two-parent family, five children, one grandchild	Elder female sibling caring for three younger siblings and their two children	Elder sibling caring for her son and three children of a younger sibling
	Mrs N belongs to three stokvels, which are very important financial and social resource	Membership in one stokvel	Membership in savings stokvel	Zionist church membership important to family well-being
	Family ‘orthodox’ church membership important to family well-being	Social agricultural network important source of support and information	Family ‘orthodox’ church membership important to well-being	Dependent on civic association for part-time road-maintenance work
		Zionist church membership important to family well-being	Family income severely reduced due to the activities of the civic association
Financial capital	Steady income from Mr N's building contract work	Income from Mr K's full-time factory employment	Income from disability grant	Income from childcare grant
	Income from stokvel activities	Income from Mrs K's baking and selling food in the community		Income from household head's part-time road maintenance work
	Savings kept in a stokvel	Family keep cash and savings in a bank
Human capital	Mrs N has a matriculation certificate and an unspecified post-school qualification	Mrs K has a Grade 8 education	Household head has some junior school education	Household head has some junior school education
	Mr N has Grade 10 and skills in building and electrical work	Mr K has a Grade 9 education	Brother disabled and unable to work or help with any chores	Brother diabetic and unable to work
				Loss of income due to death of a sibling who was sending remittances
Physical capital	Brick house	Small brick house in poor condition	Small brick house in very poor condition, leaking	Small wattle and daub house in very poor condition, leaking
	Electricity supply	Electricity supply	Electricity supply	No electricity
	Piped water	Water tap in garden	Water tap in garden	No tapped water available
	Flushing toilets	Unfenced land of sufficient size and quality to grow vegetables for year-round consumption	Unfenced land of sufficient size and quality to grow some vegetables in summer only	No land suitable for agriculture
	Fencing around garden
	Fenced land of sufficient size and quality to grow vegetables for year-round consumption

Similarly, community leadership affects the four households differently. A key community-level leadership body, the civic association, has had both positive and negative food security influences among households in the case-study community. Household Three's income has on the one hand been severely constrained by their alleged deliberate exclusion from community projects, while on the other hand the civic association has facilitated critical part-time employment for a member of Household Four (Table 5).

4.2 Inclusionary and exclusionary social capital outcomes

The cases above highlight two significant constraints in accessing a valuable form of social capital in the case-study community. The first is the fact that access to a key form of household social capital resource (stokvels) is conditional upon a household having the financial resources to make the required contributions. This suggests that the condition of being resource-poor or capital-constrained may itself lead to exclusion from accessing social capital resources. This finding is echoed in a number of studies showing that such social capital cannot be drawn on by the poorest in a community, which serves to further marginalise them (Dikito-Wachtmeister, 2001; Markussen, 2002; Cleaver, 2003).

The second is the role of specific relationships between individuals, and the role of community institutions and leadership in shaping power dynamics within the community. Despite social networks reportedly being extensively drawn on in the community in times of economic shock (Table 5), relational dynamics evidently lead to both inclusionary and exclusionary social capital outcomes. In some instances, resource poverty in households may actually be caused or reinforced by community groups and institutions rather than alleviated.

4.3 Social capital outcomes and community disunity

This section discusses three factors related to community conflict and disunity, which function to either create or reinforce inequalities and power imbalances between individuals and groups: (1) ambiguous and weak community leadership (characterised by conflict), (2) the breakdown of family units, and (3) the development of negative community norms (norms seen by informants as having a negative effect on food security and well-being).

Conflict between tribal and government leaders is not unusual in South African communities (Chimbuya et al., 2004). A number of key informants identified conflict between the traditional leader and the civic association, both of whom claimed community authority in interviews. The split between followers of Zionism and more orthodox Christian groups was also identified as a source of general divergence among community members. Previous initiatives, whether externally or internally driven, appear to have seldom been viewed by community members or leaders as politically or religiously neutral, and support for these has been bifurcated according to affiliations. The failure of an informal income-generating sewing project, for example, was viewed by informants as directly attributable to power struggles between the civic association and the traditional leader.

In the literature on leadership, management and development, unresolved conflict in communities is recognised as having a number of negative impacts on people's lives (Dimock, 1987; Fisher, 1993; Ife, 1995; Weyers, 1998; Goleman, 2003). It may, for example, lead to distorted data being offered or decisions being taken by community members or leaders to protect or promote self-interest, or to no decisions and action being taken for the community's benefit because parties are locked in power struggles that leave no space for forward-looking leadership or problem-solving (Ife, 1995; Mzwanele, 1996; Goleman, 2003). Conflict undoubtedly inhibits the development and maintenance of trust between community members, and trust is understood to be the underlying mechanism that reduces the cost of beneficial social exchanges (Barr, 2004). Ambiguous and conflicted leadership has also been found to inhibit strong connections between community leaders and external institutions, thereby stunting any potential benefits to households that might accrue from this kind of bridging social capital (Dimock, 1987; Adger, 2003; Krishna, 2003; Pretty, 2003).

Community leadership and the potential for bridging social capital have reportedly been further weakened by the ward councillor's failure to get involved in the community, which key informants attributed to his political party affiliation. Ward councillors in South Africa ostensibly represent everyone living in their ward and they have the legislative power (on paper) to open channels of communication and government resource availability (Department for Provincial and Local Government, 2003).

The impact of any particular social capital attribute on household livelihoods has been argued to be at least partly household specific. However, the extent to which a social capital attribute in the community either supports the development of further resources or erodes the existing ones may hinge to some extent on whether the attribute causes or increases conflict in the community.

The second factor related to conflict and disunity is the decline, perceived by the informants, in two-parent families and the increase in the number of orphans. Death of household members was reported as a recent livelihood shock in 42 per cent of the surveyed households, being by far the most common economic shock experienced. This figure probably reflects, at least in part, the extensive impact of HIV/AIDS identified by informants and indicated by the high infection rates in KwaZulu-Natal (HSRC, 2005) (Table 6). While the survey offers no time-series data, well over one-half of the households surveyed (62 per cent) were found to be lacking one or both parents. Some households were headed by a single adult who cared for the children or grandchildren, or both, and some consisted only of children caring for each other.

Table 6: Shocks to household income experienced by surveyed households, in descending order of frequency of citations

Type of shock	Number of households
Death of household member or other family member	21
Serious illness or injury keeping household member from normal activities	9
Loss of regular job of household member	3
Theft, fire or destruction of household property	3
Cut-off or decrease in remittances to household	2
Abandonment or divorce	1
Cut-off of government grant(s)	0
Failure or bankruptcy of business	0
Other	0

The views expressed in the four in-depth household interviews and the views of the traditional leader show that the family unit itself is seen as a valuable resource in the pursuit of food security and well-being, as well as in the maintenance of cultural norms that lead to positive food security outcomes (Table 4). In support of these community views, the literature shows that close family ties are important social ‘glue’ in the production of bonding social capital, being the primary site for generating norms of trust and cooperation, from which basis wider forms of social capital can be created (Coleman, 1988). Marriage relationships can, for example, be an important factor determining household earnings, as well as a woman's community participation and empowerment (Parcel & Menaghan, 1993; Dikito-Wachtmeister, 2001; Shinns & Lyne, 2004).

Among the specific changes in cultural norms (the third conflict and disunity factor) identified in interviews was increased alcohol abuse by adults in the community, which interviewees associated with a loss of family values and traditions and saw as an indirect cause of food insecurity. Two key informants also cited drug abuse among the children at the high school as a threat not only to the children's prospects but also to the well-being of the community as a whole.

Key informants also noted that people increasingly expect formal employment and are becoming less willing to engage in agriculture or seek alternative income-generating activities. Counter-intuitively, the proximity of the case-study community to urban centres was not found to have stimulated informal links between household livelihoods and these potential markets. The low level of informal income-generating activities and the perceived move away from agriculture may be partly accounted for by the lack of access to the minimum financial capital required to get started, as perceived by the informants (Table 7). This paper proposes that the community's expectations of employment-based livelihoods (arguably increased by urban proximity) are thwarted partly by insufficient human capital (skills) and bridging social capital (contacts) to link people securely to local labour markets. The negative impacts on food security of increased dependency on formal employment noted by informants (such as disincentives for home food production) have been noted in other food security research in South Africa (Hendriks, 2005).

Table 7: Reasons given by households for not engaging in agriculture

Reasons for not engaging in any/more agriculture	Number of times selected
Not enough time	24
Not enough water	19
Other – poor quality of soil	13
Other – no fencing	12
Not enough land	9
Lack of physical ability to do the work required	9
Seed not available	8
Poor quality of crops produced	7
Do not know enough about growing crops	7
No money for seed or fertiliser	6

5. IS A FOCUS ON SOCIAL CAPITAL USEFUL?

5.1 Implications for interventions

The case study suggests, firstly, a broader lesson for food study interventions: the need to acknowledge that food security is not interpreted in the same way across all contexts. Furthermore, because there are many facets to the food insecurity issue, there are likewise many initiatives that aim to alleviate food insecurity by many methods and means. These range, for example, from health and nutrition interventions such as supplementary feeding programmes, to early warning systems such as FEWSNET, the US Agency for International Development's Famine Early Warning System Network. Interventions also take place across geographic scales of targeting, and across scales of decision-making, from local to national policy frameworks.

‘Social capital’ is perhaps an even more complex and fraught term than ‘food security’. It is a label given to a whole suite of social attributes, and our empirical understanding of the way these interact and they way they function as mediators in the availability, access and use of other resources is not extensively developed. Moreover, interviews in the case-study community suggest (unsurprisingly) that social capital is not readily conceptualised by community members in the same way as it is understood by those researching it. This further complicates research designed to understand social capital and the development of interventions intended to enhance it. The findings in this case study, however, suggest that while social capital does not offer a panacea for food insecurity, there is value in focusing on social capital in the development of food security interventions.

The community interviews suggest that negative attributes of household-level social capital are indicative of certain risks and vulnerabilities. The evidence discussed above suggests that food security in the case study may be linked to the presence of two or more parents in a household, together with the potential to build social capital that is found elsewhere to be associated with close family ties (Coleman, 1988; Parcel & Menaghan, 1993; Dikito-Wachtmeister, 2001). Further, divergences between religious groups, conflict between leadership figures and changes in cultural norms are considered to deplete bonding (community cohesion) and bridging social capital in the community and to heighten food insecurity. These hindrances are compounded by the apparent absence of a community representative in government institutions who might voice community needs and develop collective action. Clearly, in the context of the case-study community, interventions that can resolve at least some of the leadership difficulties in the community are likely to improve household food security.

Secondly, the case study finds that social capital has an impact on vulnerability and that food security proxies for social capital arguably remain an important broad indicator of community vulnerability. The inclusion of proxies for social capital in vulnerability indices and in food-insecurity mapping systems might therefore make these systems of measurement more robust. Proxies might include factors such as trust and membership in groups, which are currently the most commonly included. The case study suggests others that need to be added: community perceptions of leadership effectiveness, conflict levels in a community, and the extent and strength of links to key government resources (e.g. ward councillors and agricultural extension) – particularly in the context of South Africa.

Thirdly, the case study emphasises that social capital is not a resource that can be conceptualised independently of other capitals, or of the context in which it operates. There is a high degree of inter-household variability in household social capital assets and the outcomes in the form of well-being accruing from community-level social capital assets. Social capital resources take on full meaning only when it is understood how people use their attributes to their advantage in the context of a community or household. In planning and developing community interventions, it is thus critical to explore the social capital context of beneficiary communities. Broader indicators of social capital resources, such as those explored through indices and mapping systems, should be interpreted only in conjunction with ongoing civic participation, in order to fully understand the nature of the groups, networks, individuals and power dynamics that shape food security.

Social capital commentators from both within and outside the region have reflected further on this need to enhance civic participation in decision-making and intervention processes (Sirianni & Friedland, 1995; Portes & Landolt, 1996; Dikito-Wachtmeister, 2001; Markussen, 2002; Cleaver, 2003; Pelling & High, 2005). Key questions might include: What key religious groups are there? What are the roles of stokvel groups, and what constraints are there to belonging to them? Is there more than one leadership figure, and does this cause conflict or weaken collective action and civic participation in development processes? Precisely because it demands a contextual approach, investigating food insecurity through the ‘lens’ of social capital calls for a consideration of the internal as well as external socio-political conditions of food insecurity (see e.g. Falk & Harrison, 1998; Gittel & Vidal, 1998; Daubon & Saunders, 2000; Narayan & Cassidy, 2001). Given food security's multiple dimensions, food security interventions aimed at, for example, helping people grow more food must also understand the social context that enables them to do this effectively. Which groups or individuals mediate household access to land and other resources? Is credit availability a prerequisite and what institutions control this? It is critical to include social capital in community ‘baseline’ assessments prior to the development of interventions. Interventions that fail to consider the social capital dynamics of their target beneficiaries are likely to have, at best, limited beneficial outcomes. This point is well illustrated by lessons learnt from the pilot study of the Food Insecurity and Vulnerability Mapping Systems (FIVIMS) in South Africa. These lessons point to institutional failures such as inadequate engagement of all user groups among government departments (Department of Agriculture and Land Affairs, 2005), poor inter-departmental and intra-departmental relations, the need for better coordination at various levels of government, and the need for greater political will (FIVIMS-ZA, 2005).

A number of questions prompted by this case study, or simply left unanswered, are worth reflecting on briefly here. Transformations in governmental, social and cultural systems (such as democratisation in South Africa, and economic development processes more broadly) bring continuing changes in livelihood dynamics and the factors that shape food security. Do households in developed communities rely on the same forms of social capital as households in economically undeveloped or developing communities? Do, for example, financial capital (usually in the form of employment), the ability to command bridging capital (in the form of relationships with people in positions of power), and market-based relationships (such as crèche child care rather than family childcare) become more important forms of capital in the livelihoods of people living in developed areas than in the livelihoods of those in developing areas? Do such forms of capital ‘replace’ to some extent extensive family or other horizontal or bonding social networks and institutions? Should change agents consider communities as ‘communities’ at all, or rather as geographically proximate but disconnected individuals in the South African marketplace?

These questions clearly require further research and analysis in South Africa and southern Africa if the relevance of social capital to food security interventions is to be more effectively explored. They also emphasise the critical need noted above for civic participation in decision-making and intervention processes in South Africa.

5.2 Conclusions

The complexity of capital resource use and interconnectedness in the case-study community makes it a challenge to isolate social capital as a food security resource. The highly variable, mutable and household-specific nature of social capital's role in community food security makes it difficult to extrapolate generic lessons for targeting social capital in food security interventions. The difficulty of using a social capital focus is further compounded by the abstract and intangible nature of the concept itself.

Within the inevitable limitations of a small case study examining a broad and complex issue, this paper has argued that, since social capital is a significant dimension of well-being and food security, there is value in including proxies for social capital in vulnerability indices and food-insecurity mapping systems. Perhaps more importantly, in its attempt to disengage social capital in attributing food insecurity causation and in considering social capital for food security interventions, the study emphasises the importance of pursuing a nuanced understanding of the interactions between various resources or ‘capitals’, and of the human relationships and power dynamics that shape food security and the outcomes of food security interventions.