Understanding rural livelihoods in the West Coast District, South Africa

Abstract

This article applies the sustainable rural livelihoods approach to purposefully collect and analyse data on how the rural poor live and work in South Africa's West Coast District. The findings offer insights into the livelihoods and needs of rural households and offer vital lessons for pro-poor agrarian reform and rural development policies. Agricultural and non-farm rural households rely on a mix of livelihood sources such as wage employment, agricultural activities and social grants. Livelihood assets are unequally distributed among farmers, farm workers and non-farm rural households. Whereas some land reform beneficiary farmers appear to be accumulating wealth (land, livestock and some financial capital), workers dependent on local agricultural labour markets are trapped in asset poverty. Effective agrarian policies should be grounded in a solid understanding of the land-based livelihood strategies and aspirations of the rural poor.

Keywords:

• sustainable rural livelihoods approach
• rural development
• agrarian policy
• civil society organisation

1. Introduction

The central objective of this study was to gain a deeper understanding of rural livelihoods in the West Coast District of South Africa. Against the background of the food and energy price increases that took place in mid-2008 (NAMC, 2009), the study also sought to gain a better understanding of the strategies rural households use to cope with such economic shocks.

Numerous local case studies report evidence of changes in the way rural South African households make a living (Francis, 2002; Bryceson, 2004; James et al., 2005; Ewert & Du Toit, 2005; Eastwood et al., 2006; Palmer & Sender, 2006). Fewer rural households currently depend on land and natural resource assets to work and survive than was the case 50 years ago. Almost two decades after the end of apartheid, redistributive land reforms have neither reversed this trend nor secured sustainable livelihoods for resource-poor farmers and farm workers (Aliber & Hart, 2009). Examining South African evidence in the context of global experiences, Bryceson (2004) argues that a similar decline in farm-based livelihoods has also been observed in underdeveloped countries in Asia and Latin America, albeit at a slower or faster pace.

Livelihood strategies in the West Coast District of the Western Cape Province are neither well understood nor well documented. Reports by the Surplus People Project (SPP), an NGO actively working on land and livelihoods in the area, sketch an incomplete picture of rural households struggling to survive through various on-farm and off-farm livelihood strategies (SPP, 2009). Other reports, such as those produced by the provincial government, fail to provide a comprehensive picture. On the contrary, provincial government reports that compare living standards across districts within the province portray the West Coast as having the lowest poverty rate outside the Cape Metropolitan Area (CMA), the urban zone surrounding Cape Town. This article explores more nuanced data to confirm or counter these accounts of living conditions and standards. Interventions intended to achieve rural transformation must start with a better understanding of the livelihood strategies and aspirations of those living and working in the agrarian sector.

This study sought to answer three basic questions about rural livelihoods in the West Coast District. First, who are the rural poor in this district and what are their livelihood strategies? Second, what kind of productive assets (land, livestock etc.) do households in the West Coast own and how are these assets accessed and used? Third, what are the land needs in rural parts of this region?

Each question is examined separately to reflect on salient insights derived from purposefully collected household data. The presentation of this set of evidence is an example of how the sustainable rural livelihoods framework might be applied to guide overall research design and data collection for specific case studies. To this end, a brief introduction to rural livelihoods analysis is sketched in the next section. We provide working definitions of sustainable rural livelihoods and key concepts that underpin this approach and also highlight relevant lessons from other empirical studies for the research. The concluding section of the paper considers the implications of the findings for the state-directed land and agrarian reforms implemented since 1994.

2. Sustainable rural livelihoods: An overview

Any description or analysis of how rural people make a living must look at these livelihoods from a logically tight conceptual perspective. The sustainable rural livelihoods approach offers one such framework. Numerous studies critically compare the strengths and limitations of this approach with neoclassical approaches to understanding the drivers and dynamics of human well-being (Ellis, 2000; De Haan & Zoomers, 2005; Scoones, 2009). This brief overview concentrates on the application of the sustainable rural livelihoods approach to this case study, particularly the way it has framed the collection of field evidence reported below.

With its origins in hybrid fields such as development studies and gender studies in the 1990s, the sustainable rural livelihoods approach does not fit neatly inside the boundaries of any traditional academic discipline. More commonly, livelihoods approaches draw inspiration from various social and natural sciences and have, in turn, provided conceptual guidance to development scholars working within and across different fields. This evolving multidisciplinary approach to socioeconomic development has gained widespread influence, enabling it to push development policy discourse, thinking and action to new frontiers (Ellis & Biggs, 2001; Scoones 2009). As an actor-oriented framework (De Haan & Zoomers, 2005), it starts from actual rural realities and focuses on ways in which rural people strive, particularly through grassroots initiatives rather than top-down interventions from the top of the policymaking hierarchy, to better their living conditions. As a practical research tool, the livelihoods approach is based in participatory action research and combines easily adaptable methods of information collection to build nuanced pictures of complex rural realities.

Before briefly reviewing the analytical concepts and tools commonly used in rural livelihoods research, we present some working definitions. What dominates the rural livelihoods research agenda is how social actors weave together varied resources and strategies to overcome deprivation. It focuses on what people do to sustain their well-being in the long run rather than just relying for their survival on safety nets that may fail them in a crisis.

Material or productive resources, codified in livelihoods research as various categories of assets, are primary determinants of human well-being. Hence, access to and control over assets receives priority attention in livelihoods studies (De Haan & Zoomers, 2005). Five types of productive resource on which livelihoods depend are natural assets, physical assets, human capital, financial assets and social capital (Ellis, 2000; Ansoms & McKay, 2010). These are referred to as the ‘capitals pentagon’.1 The pentagon displays the relative ordering of assets rather than metric measurements or estimates of their quantities or monetary values. Even in this ranking of livelihood assets, there is scope to analyse the ways in which assets may complement each other as well as any trade-offs and substitutions which may take place. Non-material dimensions of making a living now receive greater prominence in studies of rural livelihoods (De Haan & Zoomers, 2005). These include institutions and power relations, considered by many to be intangible forces, yet these forces can have a lasting tangible effect on the livelihoods people are able to construct for themselves. These concepts draw attention to processes and mechanisms that regulate or govern how people access and use productive resources and inputs – a dimension generally absent from statistical estimates of money metric poverty.

Scoones usefully defines a livelihood as an input-output process in constant interaction with its surroundings (2009:177). People change their livelihood strategies over time as they adapt in response to changes in their surroundings or as part of longer-run systemic restructuring. Ongoing research on ‘multiple livelihoods’ and ‘livelihoods diversification’ maps these wide-ranging trends, showing how rural households mix agricultural and non-agricultural activities or migrate to urban localities (Ellis, 2006). They may adjust the composition of their basket of activities to make a living, or move to locations they think will offer opportunities for better living standards. But people's livelihood activities rarely if ever take place without environmental spillover effects. The notion of ‘sustainability’, now an integral element of the livelihoods framework, captures the dynamic and complex interactions between people and their natural resource base (Hopwood et al., 2005; Scoones, 2009). Moreover, the goal of sustainability reflects the growing ecological challenges associated with contemporary rural and urban development pathways.

The challenge for researchers is to translate this multidimensional conceptual apparatus into a tool for real-life research with policy outcomes that raise the living standards of the poor. The first methodological step is to collect relevant and useful information on asset profiles of individuals, households or communities; institutional and power relations governing resource control and access; and ecological sustainability. The second step is to map and analyse evidence to inform grassroots actions and bring about visible improvements in human well-being. Ellis (2000) shows that typical rural livelihoods studies are localised case studies that make extensive use of participatory research techniques to construct household asset profiles and map social relations that affect access to resources. Using largely quantitative empirical methods, Ansoms & MacKay (2010) assembled a representative sample of people living in rural Rwanda from two national datasets: a household survey and food security data. From the derived dataset, they constructed seven multidimensional livelihood profiles ranging from rural entrepreneurs to female-headed households. Almost all other households were resource-poor (as evidenced by their small landholdings and few livestock) and often trapped in remote locations with inadequate public services, even where the soil was fertile for crop farming. Ansoms & MacKay (2010) recommend broad-based agricultural development policies to help Rwandan peasants diversify their livelihoods away from an almost total dependence on subsistence crop farming in order to intensify their engagement with off-farm activities (including labour-intensive public works).

Our study used the traditional case study methodology – mixing qualitative and quantitative data collection techniques. As a guide to the exploratory focus group sessions and to structure the survey questionnaire, we created the following broad categories of questions:

•	Profiling the rural poor: Who are the rural poor (demographics such as gender, age, race, class)?
•	Kinds of rural assets: What assets exist in rural areas (natural resources such as land, water, forests, livestock)?
•	Asset access and ownership: How do rural people gain access to assets? How do institutional rules affect the ownership of and access to assets? How are assets allocated?
•	Barriers to access and use of assets: What are the factors (such as markets or governments) inhibiting access to assets and their use?

We used purposeful sampling combined with snowball techniques as a practical strategy to obtain our sample of respondent households. For relevant theoretical principles of this methodology for in-depth case studies of respondents usually not represented in large-scale random sample surveys because they might be difficult to locate, see Heckathorn (1997) and Noy (2008).

3. Profiling West Coast rural livelihoods

The West Coast District of the Western Cape is a diverse and changing geographic space. Many generations of coastal communities in the region have been making a living from fishing. But traditions of small-scale fishing now face a sustainability crisis, partly a result of policies to privatise the exploitation of marine resources and the actions of powerful fishing corporations. For inland households further away from settlements along the Atlantic coastline, crop and livestock farming remain vital livelihood activities. This article reports on the findings of a survey of more than 600 households drawn from three municipalities (Matzikama, Cederberg and Bergriver) in the West Coast District to attempt to answer the questions listed above.

Households were classified into three groups based on self-identification of the primary occupation of the head of household: farmers, farm workers and non-farm households. Drawing on Statistics South Africa's 2000 Income and Expenditure Survey, Palmer & Sender (2006) use a household's reported volume of farm output to construct three rural poor categories: non-farming, subsistence producers and surplus producers. In their detailed analysis of rural livelihoods in Limpopo Province, Eastwood et al. (2006) categorised households according to their dominant income source. They found that households tend to ‘specialise’, either in local agricultural-based activities or reliance on ‘external transfer incomes’, while continuing to draw on a diverse range of supplementary income sources. One insight for the present study, for example, is that the non-farm category (which includes mine workers, fisher folk, old age pensioners and unemployed household heads) might also be engaged in farming for their own account. Synthesising a large body of available evidence, Shackleton et al. (2001) found that rural households rely on a more diverse livelihood base than is taken account of in national household surveys.

Mindful of these methodological and classification insights, it must be noted that resource-poor farmers in the survey, whether they are beneficiaries of land reform or not, are very different from a livelihoods analysis point of view. Farm workers depend on wage income, with the majority classifying themselves as permanently employed and more likely to work on large commercial farms. Resource-poor farmers rely largely on their own labour rather than on hiring non-family wage workers.

Figure 1: Map of the Cape West Coast, South Africa

Table 1 summarises the basic profiles of households interviewed across the three municipal study areas. More than 80% of the households in the sample identified themselves either as farm workers or as farmers – beneficiaries of land reform (LR) or non-beneficiaries of land reform (NLR). The livelihoods of the large majority of households are therefore directly dependent on agriculture. The dominant type of livelihood across the survey sites is ‘farm worker category’, and in Bergriver more than half of the sampled households fell into this category. The ‘farmer category’, especially NLR farmers (55%), is concentrated in Matzikama, which lies on the far northern edge of the West Coast (see Figure 1), bordering the Northern Cape province. Both LR and NLR farmers are resource-poor, and their farms are characterised by a general lack of commercially oriented farming activities; that is, they seldom manage to produce beyond subsistence level.

Table 1: Household profiles, location, composition and education, by self-identified category

	Valid observations	Land reform farmers (n = 72)	Non-land reform farmers (n = 89)	Farm workers (n = 361)	Non-farm^a (n = 113)
Municipal district
Bergriver (%)	240	9.72	26.97	52.63	16.81
Cederberg (%)	153	41.67	17.98	16.34	42.48
Matzikama (%)	242	48.61	55.06	31.02	40.71
Household head gender and average household size
Male head (%)	380	65.28	62.92	52.08	78.76
Female head (%)	255	34.72	37.08	47.92	21.24
Average household size		3.17	4.08	3.91	3.63
Household head highest education
No schooling (%)	34	5.56	12.36	4.16	3.54
Primary and ABET (%)^b	306	41.67	38.2	57.06	31.86
Incomplete high school (%)	239	43.06	37.08	34.63	44.25
Completed matriculation/G12 (%)	50	9.72	8.99	3.88	18.58
College/tertiary education (%)	6	0	3.37	0.28	1.77
Notes: ^aNon-farm households depend predominantly on mining, fishing, social grants and unemployment benefits. ^bABET = Adult Basic Education and Training. Valid observations exclude missing observations.

Information on the gender and size composition of households allows a comparison of the main demographic features of agricultural and non-farm households. Average household sizes ranged from three to four members, with no substantial or generalised differences across self-identified categories. In the rural West Coast, men are more likely than women to be household heads – a well-known characteristic of rural households embedded in longstanding economic, sociocultural and political relations. Female headship of households is highest among farm worker households (48%) and lowest among non-farm households (21.2%).

To contextualise the descriptive statistics on education reported in Table 1, we offer a few comments on the meaning and measurement of human capital. Among the large variety of indicators on human capital accumulation, the highest level of education achieved is arguably the most widely used proxy to measure a person's human capital status. Disadvantages of this indicator are that it measures knowledge and skills gained through an institutionalised system of education and does not explicitly incorporate learning-by-doing and vocational training interventions, and it fails to control for inequalities in the quality of education provided or received. Many factors affect the ability to accumulate this livelihood asset, including the direct and indirect costs of education (household-level affordability) and complementary public investment in education. Generally, educated parents tend to invest more in their children's education if they are able to afford it. In other words, past and future investments in human capital are strongly correlated (Francis, 2002). Special surveys on education and skills development usually collect details on children's education but this falls outside the scope of the present study. However, despite the limitations of the self-reported level of the household head's education as a measure of human capital, it is perhaps the least onerous way of collecting relevant information in a survey not specifically designed to track human capital accumulation.

Levels of education are lowest among NLR farmers, with 12% of heads reporting that they had never received any schooling. Alongside relatively low levels of education in this sample, it is worth noting that between 75% and 91% of all household heads had attended primary and high school. Very few respondents had completed Grade 12 and the proportion with tertiary education and training was very small. Focusing on household heads with Grade 12, it is interesting to note that the share of non-farm rural households with this level of schooling was almost 19% – double the proportion among farmers and more than four times the proportion of farm workers with this qualification. Substantially higher levels of education are required if human capital is to become a pathway to sustainable livelihoods for rural households in a rapidly modernising economy.

In summary, the only significant variation in the demographic profiles of the sampled households is that levels of educational attainment for non-agricultural households are higher than for those primarily involved in farming or farm work. As noted above, the dominance of male headship and marginal differences in household sizes are unsurprising for rural settings.

4. Ownership, access and use of productive assets

The livelihood status of households and their ability to withstand shocks (sudden events with significant livelihood impacts) and stresses (slowly unfolding long-term stressors with potentially significant cumulative impacts) are largely dictated by their access to and control over assets (Ellis, 2000). Asset deprivation is a fundamental driver of households' long-term vulnerability to becoming trapped in poverty. Details for three categories of livelihood assets are reported in Table 2 – physical, financial and agricultural.

Table 2: Livelihood asset profile by rural household category

	Valid observations	Land reform farmers (n = 72)	Non-land reform farmers (n = 89)	Farm workers (n = 361)	Non-farm^a (n = 113)
Immovable assets
Brick house dwelling (%)	533	77.78	82.02	85.04	85.84
Shack dwelling (%)	39	12.5	6.74	5.82	2.65
Movable assets
Sedan motorcar (%)	91	25	14.61	9.42	23.21
Pickup bakkie (small utility truck) (%)	66	19.44	21.35	4.99	13.39
Tractor (%)	14	4.17	8.99	0.55	0.89
Financial assets (average value in rands per household)
Gross livestock sales income (per year)^b	119	8771.70	2436.90	377.34	300.00
Gross crop sales income (per year)^b	258	614.66	4868.10	143.08	519.21
Labour income (per month)	361	1673.75	2931.20	664.76	3174.41
Social grants (per month)	253	805.10	682.70	151.59	95.38
Other financial capital (per year)	223	2757.44	312.50	211.64	3511.11
Farmland assets
Farmland ownership (%)	88	38.89	49.44	3.05	4.42
Land size accessed (average hectares)		24.2	21.4	0.15	0.15
Livestock assets (average number owned per household)
Cattle	76	5	2.75	0.07	1.56
Goats	89	15.8	7.09	1.92	1
Sheep	99	51.8	22.1	1.57	1.06
Pigs	87	7.08	9.86	1.22	0.82
Chickens	75	20.5	5.57	3.23	1.13
Notes: ^aNon-farm households depend predominantly on mining, fishing, social grants and unemployment benefits. ^bEstimates exclude imputed values for household consumption, barter and donations of farm outputs. Valid observations exclude missing observations.

The distribution of physical assets as measured by types of immovable and movable assets owned or accessed is mixed. More than three quarters of the households – in the order of 77% to 86% – live in dwellings constructed with bricks and cement. A relatively small number of households reported owning any motor vehicle, which is largely indicative of the unequal distribution of wealth and the way this, in turn, determines a household's ability to invest in these assets. Ownership of these assets is lowest among farm worker households. Tractor ownership is extremely restricted and, unsurprisingly, tractors are mainly owned by farmers. Nine per cent of NLR farmers own tractors compared to 4% of LR farmers – a difference probably connected with differences between these categories in the level of participation in crop farming.

The composition of financial assets reveals that rural households survive through a mixture of livelihood sources and strategies. Financial assets are spread across multiple sources such as land-based activities (own-account and market-oriented farming), labour markets, social transfers and savings (formal and informal). This implies that even though the household head self-identifies as a farmer, the household unit could also rely on wage employment, remittances and social grants to sustain itself. Similarly, non-farm rural households and farm workers, for instance, earn incomes from the sale of farm output despite their limited access to land.

Wage labour is a major income source across all four categories but its monetary value points towards stratification and inequalities in rural labour markets. Non-farm rural households earn the highest average monthly wage of R3174, whereas the average wage for a farm worker is roughly R665. A worker in non-farm rural employment could receive almost five times more than what is paid in the agricultural labour market. Wage incomes flowing into households primarily involved in farming fall between these maximum and minimum values: NLR households earned about R2931 compared to R1673 for LR households. Given the low mean agricultural wage and similar household sizes, it is likely that farming households might be supplementing their incomes through participation in non-farm rural employment to produce higher labour incomes.

Social grant beneficiaries exist in each rural category. Although old age pensions, child support grants and disability grants are the most widely accessed, the types, number and value of grants per household differ. The average monthly values of social grants flowing to LR and NLR farmers, standing at R805 and R682 respectively, are between four and eight times higher than what other categories receive. Other financial capital includes the average accumulated values of formal and informal savings plus other forms of investment at the time of the interview. Non-farm rural households evidently accumulated the largest average value in ‘other financial capital’ (R3511), followed by LR farmers (R2757). Farm worker households are the poorest in terms of financial assets and the position of NLR farmers is only marginally better.

The farmland and livestock assets listed in Table 2 – examined alongside the revenues from livestock and crop sales – demonstrate the degree of importance attached to agriculture in rural livelihoods. As can be expected, farm households engage more extensively in agriculture than any other group of rural dwellers. When farm workers and non-farm rural households practise own-account agriculture, they do it on a small and modest scale. Only 88 households in the sample reported ownership of farmland, compared to 238 households with access to land for some farming activities. Self-reported ownership of farmland is evidently much higher among NLR farmers (50%) compared with LR farmers (39%). When it comes to the total size of land owned and accessed, LR farmers control larger amounts of ‘land wealth’ than other rural categories.

A module in the questionnaire probed actual land use for planting crops and livestock grazing in the 12 months preceding the survey (July/August 2008). On average, LR farmers used 3.75 ha for livestock grazing and more than 4 ha for crops, compared to 2.22 ha and less than 3 ha respectively for NLR farmers. Taking actual land use as a share of the aggregate amount of farmland owned and used, NLR farmers used about 25% of their land compared to 33% among LR farmers. Farmland ownership, access and use were almost invisible among farm workers and non-farm rural households, with only a few households being exceptions.

Farmers keep mainly small livestock, particularly sheep and goats, as these animals are better adapted to the semi-arid agro-ecological conditions of the West Coast. Livestock ownership is generally poor among farm workers and non-farm rural households and this appears to be mainly due to their limited control over farmland assets. Except for pig ownership, the average LR farmer owns roughly double the number of cattle, goats and sheep owned by the average NLR farmer.

Land ownership plays a critical role in sustaining rural livelihoods. It forms the productive base for crop and livestock farming, making it probably the most important production resource. The average farmland ownership for the entire sample is roughly 14% of households. Such ownership is usually verifiable with the aid of a formally registered title deed, which generally confers private ownership in terms of the current land tenure system. Within a private land tenure regime, however, it is not uncommon to find other forms of landholding and access. Moreover, resource-poor small farmers might enjoy use rights over land (irrespective of the form and level of payment) without formal title. Documenting the dozens of land tenure arrangements is a complex undertaking which would require extensive research.

The survey also included questions about land access for horticulture (orchards), but found that an insignificant minority of LR farmers (< 6%) reported having access to land for this purpose. Although they work in the agricultural sector (especially on horticultural farms), farm workers appear not to engage in either crop or livestock farming on a substantial scale. Approximately 18% of farm workers reported that they engaged in home gardening.

To sum up: in the rural West Coast, agricultural and non-farm households depend on a diverse mix of livelihood sources. However, the ownership, access and use of productive assets are unequally distributed across and within groups of farmers, farm workers and non-farm rural dwellers. Among those primarily dependent on agriculture, LR households have the most secure livelihoods base. For each class of livelihoods assets, farm workers are the most deprived rural category. Rural households with greater asset security tend to be less vulnerable to socioeconomic shocks but may be better equipped to devise resilient coping strategies in crises, an issue which is discussed below.

A household's food security status is also a useful livelihood indicator. The average household in this sample spends approximately 60% of its total income on food – a relatively high food expenditure share which is generally associated with food insecurity (Altman et al., 2009; Wallinga, 2009; Fan et al., 2011). While a majority of households did not report incidences of chronic hunger, 20% indicated severe hunger and food insecurity. In sum, a combination of factors, inter alia food and energy price increases, limited livelihood strategies, low incomes and high degrees of indebtedness, all serve to indicate deeply entrenched vulnerability to poverty traps and food insecurity for rural households in the West Coast District.

The food and energy price shocks that affected the country in 2008 are a case in point. Following these shocks, a significant portion of households (35.4%) identified food price increases as a major shock or risk affecting their livelihoods. The distributional impact of food prices across the surveyed sites shows the district of Matzikama reporting the highest proportion of affected households (25.6%), followed by the Cederberg district (21.6%). About 10% of the sampled households pointed to the rise in energy prices and energy blackouts as additional shocks that had had a negative impact on them. Further, households across all districts reported a lack of coping mechanisms to deal with such risks, whether at household or community level, and this exacerbates their severe vulnerability. Because there are not many viable livelihood strategies in the West Coast District, most households are heavily dependent on wage income.

Profound lessons can be learnt from the ways in which the 2008 food and energy price shocks affected rural households (Altman et al., 2009; NAMC, 2009). First, quantitative information on food spending patterns and self-reported evidence show that rapid inflation has hurt rural households on the West Coast. Secondly, because their incomes are low, a significant proportion of households evidently became debt-trapped as they were compelled to ‘trade on debt’. Thirdly, households that depend on social grants, such as old age pensions and child grants, appear to be the worst victims of food price shocks as these grants generally do not keep pace with the escalating cost of living.

5. Assessing land needs

Assessing the demand for land is far from straightforward. The primary goal here is to gain a sense of how much land households want for different land uses, and what steps a household intends to take to satisfy its land hunger. It is widely acknowledged that the purpose of land reform is to help black households gain access to land, despite the policy's vagueness on the wealth status of beneficiaries (Hall, 2009). It is vital to know a household's awareness of, and participation in, different land and agricultural reform programmes, because these programmes are the main policy instruments available to help resource-poor rural households gain access to land. Firstly, among rural households along the West Coast, land demand appears to be strongly split between those wanting land for agriculture and those wanting it for housing. The demand for housing land appears to be higher among farm workers, mine workers and fisher folk than among the farmers. This suggests that housing tenure security is a major concern for these sections of rural dwellers.

Turning to demand for agricultural land, interesting insights emerge on land size requirements, land tenure preferences and what households intend to do with their agricultural output. Looked at on its own, the demand for agricultural land is relatively small, yet it is, surprisingly, higher than the demand for housing land. As can be expected, the demand for farmland is concentrated among farmers, although it is important to note that nearly 25% of all farm workers wanted land for crop farming. Overall, NLR farmers expressed the strongest demand for land, with close to 50% wanting land for livestock grazing and close to 40% for crop farming. Moreover, the strong demand for land is further articulated in the average land sizes they would like to have: about 21.5 ha for crop farming and more than 50 ha for livestock grazing. This means that farmers could require a tenfold increase in the average amount of land currently available to them. Freehold title is the strongly preferred land tenure arrangement among households that articulated a demand for farmland. It is instructive to note that households planned to sell their surplus output in local markets, instead of national and international markets with much larger purchasing power but perhaps more aggressive supply-side competition.

To satisfy their land need and acquire private ownership of farmland, households rely almost exclusively on their local municipality and the Department of Rural Development and Land Reform. This dependence on the state is probably related to the present land tenure status of most farmers (who occupy land with the permission of some state agency) and poor knowledge of land reform policies and processes. What is puzzling, though, is that rural households lack a clear understanding of the workings of specific land and agricultural policy instruments. For instance, rural dwellers tend to connect their demand for land for agricultural production to land restitution (and not the Settlement/Land Acquisition Grant – SLAG, the Land Reform for Agricultural Development programme – LRAD, or the Comprehensive Agricultural Support Programme – CASP). However, since its inception land restitution has prioritised monetary compensation for land rights lost under apartheid rather than genuine agrarian development. Thus far, this rights-based element of land reform has remained a weak instrument for settling beneficiaries on productive farmland, upscaling ecologically sustainable agro-food production, and restructuring production and social relations in the agrarian sector.

6. Concluding summary and policy implications

This study has presented an illustrative example of how the sustainable rural livelihoods approach might be applied to generate and analyse policy insights from a specific dataset. Livelihoods frameworks are mainly concerned with tangible and intangible assets (social networks, sociopolitical relationships, institutions and so on) that shape the way households and communities work and live. This approach has been applied to track ongoing changes in rural areas of developing countries, including South Africa. It highlights a major shift: a decreasing reliance on land and natural resource based livelihoods and increasing livelihood diversification. Findings from the 2008 West Coast dataset largely confirm this trend.

Unlike many other studies, this study combined objective quantitative indicators and subjective self-identified categories, as suggested by Bryceson (2004), to profile several agricultural and non-farm rural households. Within each rural category, households depend on a diverse mix of livelihood sources but inequalities exist in the control of key assets. Asset inequality in rural contexts, according to Eastwood et al. (2006), drives insecure and unsustainable livelihoods. The reason for this is that output and employment in the farm sector fall at a much faster rate than the absorption capacity of rural non-farm sectors. To counter the resulting rural dependency, redistributive farmland reforms with appropriate public support for smallholder farming might be a necessary, but probably insufficient, intervention. Smoothing the transition to sustainable rural livelihoods is bound to be too complex for any one-size-fits-all solution.

This study's descriptive analysis revealed highly stratified control over livelihood assets. Almost all farm worker households, a fairly large proportion of them headed by women, are trapped in asset poverty. They have the lowest levels of education and lack the human capital they need to secure decent and resilient living standards, especially through the non-farm rural economy. Workers in non-farm rural jobs evidently earn higher mean wages but this sector is small and might be constrained by other demand-side barriers. Among farmers, on the other hand, land reform beneficiaries have been able to accumulate land, livestock and some financial assets. However, this finding does not clearly demonstrate whether South Africa's land reform is merely boosting the fortunes of already prosperous black farmers, or turning resource-poor rural producers into accumulating farmers. Better impact evaluation approaches and longitudinal data will be required to begin such an assessment.

Enabling the rural poor to build sustainable livelihoods means helping them to overcome the barriers to access and use productive assets. A first step in this process is to identify critical asset-specific obstacles to accessing education, land, finance and other productive resources. This process is mainly determined by the existence, resilience and functionality of local institutions – markets, states and social networks. Access to education, for instance, is a pathway to accumulating human capital and moving into better paying non-farm jobs, but income poverty might prevent some rural households, such as farm workers, from investing in their children's education. In this context, appropriate public investments in schooling would be vital to break intergenerational low human capital traps.

Understandably, the redistribution of farmland is one of the cornerstones of agricultural policies in post-apartheid South Africa aimed at reversing the inequities inherited from apartheid (Hall, 2009). Access to land is viewed as a critical enabler of a multiple-livelihood strategy that is particularly necessary for the rural poor. Hence it is widely acknowledged that improving resource-poor farmers' access to land constitutes a key strategy for fighting poverty. Land hunger is not really a criterion for how the state allocates resources for land and agrarian reform. Uptake and participation of resource-poor farmers in land reform is left to chance or the ‘invisible hand’ of the market. A wealth-based ranking such as the degree of landlessness, which can also be a test of the pro-poor content of fiscal spending, is absent from land reform policies. If consistently applied and revised as a result of learning from experience, land reform targets based on the land needs of resource-poor and landless farmers could improve the state's ability to achieve its vision of a more equitable agrarian structure.

Acknowledgements

The authors wish to thank the Surplus People Project (SPP), an NGO working with rural communities, for initiating and sponsoring this research, and Laetitia Louw and Stephen Heyns for their help with revising this article. In case of any remaining errors, the usual disclaimers apply.

Notes

¹ For a diagram and full explanation of the ‘capitals pentagon’ concept, see Ellis (2000:49).