The limits of land restitution: Livelihoods in three rural communities in South Africa

Abstract

Land restitution is a powerful symbol of redressing the trauma and dispossession many Africans experienced historically. However, everyday material constraints limit the possibilities of how restored land is used, resulting in continued alienation of African rural households from land. This paper focuses on three rural communities in South Africa that are recent beneficiaries of post-1994 land restitution. Social conditions in these communities create a context in which decisions on how restored land should be used, especially the tendency towards commercial farming enterprises, remain unchallenged by ordinary people.

Keywords:

• household livelihood
• land restitution
• Ebenhaeser
• Mashishimale
• Mophela

1. Introduction

A majority of black South Africans were denied access to land and the possibilities of capital accumulation as a result of land appropriation and a concomitant restrictive land policy implemented by pre-1994 governments. In the context of land restitution in a new democratic dispensation, important questions have been raised as to how the restored land is to be used. The decision as to how it should be used is almost entirely the prerogative of policy makers, in spite of the participative process implied by both the policy and the legislation. Moyo argues that ‘the peasant question (or even the small farmer development trajectory) is underestimated by official policy’ (2007:61). He contends that this is due to a presumption of inefficiency in the productive use of land in such communities. From the perspective of the state, the post-restitution process for land has a development-oriented thrust, underpinned by a neoliberal paradigm. Such an approach consequently emphasises managing restored land as an economic resource, with commercial agricultural production presented as the most efficient and sustainable use of land. Walker (2007) has noted a tension between returning the land to address past injustice and meeting the demands of agricultural production. The orientation to land restitution is underpinned by the need to demonstrate the viability of land restitution by ensuring that agricultural production is maintained on restored farms. The pressure to prove the agricultural viability of restored land is increasingly intense, particularly against the backdrop of the failure of ‘fast-track resettlement’ in Zimbabwe and the consequence of extreme food insecurity in that country (Francis, 2002; Bernstein, 2005; Tsikata, 2009; CLRDP, 2010a).

However, land restitution and commercial agriculture are two different processes, and there is no reason why the one should necessarily imply the other (Walker, 2005). While this criticism of the land restitution process remains pertinent, this paper aims to frame this debate in a way that looks beyond the constraints and motivations of policy makers. Rather, it focuses on the context of everyday life in three communities in which restored land involved a number of commercial farms. It brings the experiences of ordinary people into the frame of analysis, and shows how rural livelihoods are a further constraint on how restored land is used. Ordinary people's continuing struggles to reclaim land are at odds with the declining contribution of agricultural activities in the chain of households' viable survival strategies. This creates a context where the state's coupling of commercial agriculture to the processes of land restitution remains largely unchallenged by ordinary people. Furthermore, the expectations of the people themselves have shifted and society is no longer agrarian in its disposition or aspiration. A large proportion of people in these communities are oriented towards a consumer-based lifestyle in which a secure cash income plays a key role. Hence, acquiring employment or pursuing other more dependable means of accessing cash tend to be the preferred livelihood strategies.

Below, the literature on livelihoods is introduced briefly. Thereafter the three communities are described and a short overview of the research design underpinning the study is provided. The current socioeconomic context of these communities in which claimants have successfully applied for restitution of their dispossessed property through the transfer of full title deed is considered, followed by concluding remarks.

2. The question of rural poverty and livelihoods

Research on livelihoods gained impetus during the 1990s, in a context in which the de-agrarianisation of rural communities in Africa and the multiple links to other economic activities became increasingly evident. Rural livelihoods implied not only a diversification of economic activities but also closer and more extensive connections with urban centres. Within this context, households manage risks by diversifying their sources of livelihood (Ellis, 1998; Bryceson, 2004). Ellis and Freeman identify a ‘clear and shared preference ordering’ among these various livelihood strategies (2005:17). The shift to non-agricultural sources of livelihoods, and in particular the monetarisation of such livelihoods, continues to be evident. It is in this context that Murray observes that the paradigm of sustainable livelihoods ‘is now strongly entrenched in official discourse’ (2002:492).

How to manage resources within the limits of the natural environment while facilitating improved livelihoods in marginalised regions is therefore a key objective of government and non-governmental bodies alike. The notion of sustainable livelihoods aims to be, in a sense, all-encompassing. It considers the natural environment, physical infrastructure, social relationships, human skills and also financial assets as it gauges the degree of vulnerability or resilience of a household or community to cope with economic shocks or setbacks. However, in spite of the ever-growing trend towards non-agricultural sources and the monetarisation of livelihoods, agriculture has remained the livelihood strategy most often targeted by agencies implementing rural development projects to address poverty. Debates are also framed around the evident and worrying collapse of formal employment opportunities, which have increased the strain on many rural households. A key concern is about who ultimately makes the critical decisions to alleviate pressures on these households, bearing in mind especially the delicate interconnectedness between survival strategies, nodes of power in communities, and various webs of social relationships. Murray (2002) emphasises that there is a naive presumption that those in power make better policies for the ‘poor’. Local notions of sustainability may differ from those held nationally or by international aid agencies. Here a disjuncture emerges between the community and policy (Hajdu, 2005). Arguably such fissures undermine the viability of ‘sustainable livelihoods’ – whichever way such a notion is defined.

However, and this is what our analysis brings to the fore, we must consider how and to what degree policy makers' decisions may be compatible with what ordinary people can accommodate in their livelihood strategies in practice. Because there is too little participation by community members in the various structures involved with the land restitution process, as is the case in the three instances in this study, this ‘uneasy congruence’ is not mutually negotiated but rather imposed by survival strategies and aspirations deeply embedded in a cash economy. The complexities of everyday living in South Africa ultimately leave critical decision-making in the hands of more powerful authorities.

3. Overview of the communities

The three communities studied here are Ebenhaeser, in the Western Cape, Mashishimale, in Limpopo Province, and Mophela, in KwaZulu-Natal. The Mashishimale and Nkumbuleni land claims were legally finalised, i.e. approved by the Minister of Rural Development and Agriculture, by the time the surveys were conducted in 2008. A majority of the Nkumbuleni claimants lived in Mophela, where some land claims were unsuccessful, and a number were still pending at the time of the research (Hlongwa, 2004). The community in Ebenhaeser was awaiting the Minister's final approval (CLRDP, 2010b). No beneficiary community had settled onto restored land as yet. Restored land for both Mashishimale and Ebenhaeser bordered on land occupied by the communities.

The mission station around which the Ebenhaeser community had been established was relocated to Nuwestasie (‘new station’) in 1926. This ward is the core of greater Ebenhaeser, with a primary school, sports fields, community hall, satellite municipal office, clinic, post office and library. Dwellings emulate suburban housing in adjacent towns. The closest secondary school is at Lutzville, 18 kilometres away. To the east of Nuwestasie is a newer development, where most of the RDP houses (low-cost houses built for the community as part of the Reconstruction and Development Programme) are located, known locally as HOP-land (‘RDP-land’).1 Larger allotments for cultivation are located in the farming areas of Rooi-erwe (‘red plots’), Merriandale and both North and South Nuwepos (‘new post’). Subsistence fishermen and livestock farmers live in Olifantsdrift and Papendorp. Ebenhaeser had a total population of 1723 according to the 2001 Census figures. The number of inhabitants was projected to increase to over 4200 by 2008 based on an expected high population growth rate reported in the Matzikama Integrated Development Plan (MM, 2007).

Mashishimale village, situated within the boundaries of the Ba-Phalaborwa municipality in Limpopo, is an area under the jurisdiction of a tribal authority. The township of Namakgale adjoins it in the east and the restored farms lie along its western border. The area in which the community was instructed to settle in 1922 is known as R1. As the community grew, R2 developed on the opposite bank of a tributary of the Selati River, which runs through the area. R3, the newest development, is contiguous with R2, and closest to the main road. Offices of the traditional authority, a community hall, clinic, primary and secondary schools, churches, shops, taverns and rudimentary soccer fields are located within the community. According to Census 2001 there were 2745 homesteads and a total population of 12 391 in Mashishimale (StatsSA, 2008).

Mophela is under the jurisdiction of a tribal authority. It is located within the boundaries of the eThekwini Metropolitan Municipality. The settlement can be divided into two parts. Sankontshe is larger and is more densely populated than the smaller Mophela proper. There were approximately 7156 households and 34 122 people in greater Mophela in 2008. The Nkumbuleni claimants were evicted from the Ithala Ngomankulu Valley in 1933. These claimants have agreed not to settle on their restored farms located near Camperdown.

4. Research design

Three rural land claimant communities opting for transfer of dispossessed land back to them, on which commercial farms had been established, were purposively selected2 for a pilot project on sustainable restitution support. The selection accommodated the ‘geographic and demographic land restitution profile’ of the country, as well as ‘pre- and post-settlement’ contexts (CLRDP, 2010b:xiv, xv).

During initial visits the research team met a non-governmental organisation facilitator appointed for each community who introduced them to community leaders. The team explained the purpose of the survey, sought permission to conduct a survey and accompanying fieldwork, developed and tested the interview schedule, and used unemployed adults from the community as local enumerators, training them in the technique of interviewing and the principles underpinning ethical research, such as voluntary participation, the right to withdraw, confidentiality and privacy.

Baseline surveys were conducted between March and May 2008, using a one-stage systematic random sampling methodology. Population figures for each community, together with standard methods, were used to calculate a household sample size of 373 for Mashishimale, 380 for Mophela and 314 for Ebenhaeser at a confidence level of 95%. A sample of 379 in Mashishimale maintained a confidence interval of 5% and a confidence level of 95%. Difficulty in recruiting suitable fieldworkers within a limited time frame reduced the sample realised in Mophela to 311 households sampled, increasing the confidence interval to 5.53%. A smaller sample of 298 households was also realised in Ebenhaeser, due to technical difficulties, resulting in a confidence interval of 5.16%.

Local enumerators and Department of Sociology postgraduate students interviewed an adult responsible for key decisions in each household with regard to its demographic composition, income, expenditure and assets. Questions were also asked about infrastructure in the community, household and community needs and, in the final instance, both perceptions and expectations of land restitution. These data were supplemented by a wide range of informal conversations with key informants in the respective communities, and by available documents. The surveys provided data on living conditions in Mashishimale and Mophela shortly after land was restored to the claimant communities (and in the case of Ebenhaeser just before this process was finalised).

5. Demographic correlates

Demographic correlates affect the structure of communities and households within them, shaping the context in which livelihoods are negotiated as well as constraints on household livelihood strategies.

5.1 Age and sex structure

Table 1 reports data on the age and sex structure of these communities, as well as corresponding national data for comparison.

Table 1: Age and sex structure

	South Africa 2001	Ebenhaeser	Mashishimale	Mophela
Median age	24	29	22	23
Youth dependency ratio	51	41	54	45
Aged dependency ratio	8	16	8	6
Total dependency ratio	59	57	62	51
Sex ratio	92	98	93	81
Notes: The dependency ratios express the youth age category (0 to 14) or old aged category (over 64), or both, in relation to the economically productive category (15 to 64). The ratio reports the number in these dependent categories in relation to 100 in the economically productive category, providing an indicator of economic burden. The sex ratio reports the number of males in relation to 100 females.

The median age in both Mophela and Mashishimale, calculated from the 2008 survey data, is slightly below the national median age, which is based on 2001 Census data (StatsSA, 2003). In contrast, the median age in Ebenhaeser is at the upper end of the age range for an intermediate population age structure (Hobbs, 2004), suggesting a trajectory towards population ageing. All three communities still have a significant proportion of dependent children and, as is most clearly evident in the case of Ebenhaeser, an increasing proportion of aged persons. Although the dependency ratios shown in Table 1 are crude generalisations, they provide an indication of the potential burden for the communities, by looking at the numbers of people younger than 15 and older than 64 in relation to the potentially economically active segment of the population (15 to 64). The dependency ratios in these communities suggest a continued demand for expenditure on social services such as education and health.

Ebenhaeser, with its low level of migration, reflects the closest approximation of a balance between sexes, proportionately. The sex ratio of 93 in Mashishimale corresponds closely to the South African sex ratio of 92, calculated from data of the 2007 Community Survey, cited in RSA (2009). In contrast, the sex ratio of 81 calculated from survey data for Mophela reflects an imbalance between sexes. Although the disproportionate number of females to males has been more pronounced in rural areas because males enjoyed more mobility in the past, current employment opportunities militate against using this as an explanation for the imbalance observed in Mophela.

5.2 Household demographics

The mean household size, based on the survey data, was 3.1 for Ebenhaeser, 3.7 for Mashishimale and 4.5 for Mophela, compared to 3.9 for South Africa (RSA, 2009). The large development of RDP houses, spread across three wards within Ebenhaeser, enabled a number of young couples to obtain property of their own, thereby reducing the average household size.

The size of a household and its age and gender composition potentially affect its access to resources and therefore mould livelihood strategies (Hajdu, 2005). The nuclear family, consisting of a married or cohabiting couple and their offspring, was the most common arrangement for households in Ebenhaeser and Mashishimale. This reflects a trend towards independent self-supporting, self-contained homesteads. However, more than a third of households in both these communities were extended families. Mophela had a higher proportion of extended families, and a larger mean household size. Most extended families spanned three generations and reflected a variety of forms, such as children starting a family but not being able to establish a homestead of their own, or parents seeking employment elsewhere, leaving their children in the care of grandparents. Household composition, as Hajdu (2005) argues, is suggestive of both potential resources and vulnerability. The development cycle of a household determines both the resources that can be accessed and the economic opportunities that can be exploited. For instance, in the course of time there is a shift from wages to pensions associated with a generational shift in the structure of the household.

5.3 Household headship, age, marital status and education

Almost two thirds of the households in Ebenhaeser and three fifths of those in Mashishimale were male-headed, whereas only two fifths in Mophela were male-headed. The large number of female-headed households in Mophela reflects, in part, the larger proportion of females in the community. There was no significant age difference between male and female household heads. Proportionately more male household heads were married (see Table 2).

Table 2: Selected demographic characteristics of household heads

	Male	Female	Total
Married (%)
Ebenhaeser	66	34	100
Mashishimale	61	39	100
Mophela	39	61	100
Never married (%)
Ebenhaeser	11	34	19
Mashishimale	11	38	21
Mophela	26	51	41
Mean age (years)
Ebenhaeser	56	52	54
Mashishimale	48	52	49
Mophela	52	53	53

Establishing a homestead ensures a greater degree of autonomy over resources. For men in particular this is ideal. It also offers women an escape from patriarchal dominance. Death or desertion of a partner may leave a woman as the head of a household. Although reasons for establishing an independent household may vary, doing so successfully requires access to resources (Francis, 2000; Tsikata, 2009).

The increase in female-headed households in rural communities can be attributed to weakening family structures as a result of migrant labour, as well as better access for women in the labour market. Employment opportunities in male-dominated sectors have diminished and a wider range of social grants is accessible, enabling them to support their families and reduce their dependence on men.

Ebenhaeser had the largest percentage (54%) of households with members that have secondary schooling. This was closely followed by Mophela with 48% (see Table 3). A basic level of literacy is evident in all three communities, since only 8% in Ebenhaeser, 14% in Mophela and 18% in Mashishimale had no schooling at all. The improvement in access to education is apparent when data from the three adult age categories in the table are compared. Data in the table suggest that the younger adult age cohorts have had better access to education – that is why for instance between 0.6 and 3% of those in the 20 to 39 age group have no schooling, while the equivalent percentages for the over 65s range from 11% to 75%. The trend is also evident when attainment of some primary or secondary schooling is compared across the adult age categories.

Table 3: Level of education by age cohort

Age	No schooling			Primary			Secondary			Post-school
	Eb	Ma	Mo	Eb	Ma	Mo	Eb	Ma	Mo	Eb	Ma	Mo
0 to 19	16	20	21	56	56	50	28	26	29	0.5	0.5	0.2
20 to 39	0.6	2	3	14	14	10	83	73	80	2	12	7
40 to 64	4	31	17	31	38	47	61	26	31	3	5	3
65 +	11	75	49	46	23	33	42	2	18	0.8	0	0
Total	8	18	14	37	37	35	54	40	48	2	5	3
Notes: Eb = Ebenhaeser; Ma = Mashishimale; Mo = Mophela; TOT = proportions in level for each area. The figures, rounded, represent proportions expressed as percentages within an age category for each level in the community.

Ebenhaeser, the smallest of the three communities, did not have a secondary school. In spite of this drawback it compared favourably with the other two communities on levels of literacy attained by its age cohorts, registering the lowest proportions that had no schooling across all age cohorts, probably due to better socioeconomic conditions, historically. Primary health care needs are addressed by a clinic service in all three communities.

6. Income, expenditure and employment

A striking characteristic of these communities, in spite of their established infrastructure, is the difficult material circumstances that a significant number of households face. Households function primarily in a cash-based environment to meet their needs. Various livelihood strategies generate income. The importance of cash in the negotiation of livelihood strategies can be seen when both production (denoting work and income) and consumption (denoting expenditure and assets accrued) are considered.

6.1 Household income

Whereas proportionately more male-headed than female-headed households have at least one adult member who is employed, earning a wage, conversely, proportionately more female-headed than male-headed households have access to social grants (see later Table 6). However, in Mophela about 52% of female-headed households had a monthly income of R1000 or less. By contrast only 28% of households in Ebenhaeser, and approximately 48% of households in Mashishimale, were in a similarly precarious position, reporting a monthly income of R1000 or less (see Table 4).

Table 4: Reported household income

Category	Number of households			% of households
	Eb	Ma	Mo	Eb	Ma	Mo
R1 to 1000	83	183	162	28	48	52
R1001 to 2000	87	67	29	29	18	9
R2001 to 4000	64	32	22	22	8	7
R4001 to 6000	24	10	8	8	3	3
R6001+	20	8	4	7	2	1
No income	12	0	0	4	0	0
Not disclosed	7	79	86	2	21	28
Total	298	379	311	100	100	100
Notes: Eb = Ebenhaeser; Ma = Mashishimale; Mo = Mophela.

Table 5: Household expenditure

Item	Amount (rand)			Expenditure (%)
	Eb	Ma	Mo	Eb	Ma	Mo
Food	737	544	683	40	34	21
Education	68	222	347	4	14	10
Clothing and footwear	161	218	341	9	14	10
Toiletry, detergents and other non-food items	102	51	129	5	3	4
Energy (electricity, gas, etc.)	122	79	117	7	5	4
Transport	264	135	317	14	8	10
Hire purchase instalments	163	196	545	9	12	16
Water	71	3	121	4	0.2	4
Cell phone/telephone	84	41	95	4	3	3
Contribution to burial societies	66	92	126	3	6	4
Other	18	4	487	1	0.2	15
Total	1856	1585	3308	100	100	100
Notes: Eb = Ebenhaeser; Ma = Mashishimale; Mo = Mophela.

Table 6: Income by sex of household head

	Ebenhaeser		Mashishimale		Mophela
	Male	Female	Male	Female	Male	Female
Wage	56	37	57	47	61	53
Remittance	3	6	6	3	2	6
Social grant	61	85	70	81	64	72
Other source	11	21	15	12	5	2

To determine household income and counter memory lapses, interviewees were asked to report on their income for the month preceding the survey. They were also asked whether it was more or less than, or the same as, their income for previous months. Interviewers were trained to probe interviewees about the contributions, if any, that all adult members made to the household income. The difficulty of determining household income is demonstrated by the fact that more than one in five households surveyed in Mashishimale (21%) and Mophela (28%) declined to disclose their monthly income. Although only 2% of households in Ebenhaeser declined to disclose their income, a further 4% claimed they had no income at all. As can be seen in Table 4, only a few households reported income above R6000 per month. The highest income reported was R25 000, an exception rather than the rule. Ebenhaeser had proportionately more households in higher income brackets than the other two communities. The mean monthly household incomes reported were R2302 in Ebenhaeser, R1670 in Mashishimale and R1464 in Mophela. The median monthly household income was R1800 in Ebenhaeser, R929 in Mashishimale and R940 in Mophela.

Differentiation in income between areas within these communities was evident as well. In Ebenhaeser the mean income of the established Nuwestasie was the highest, while that of the underdeveloped and more isolated Papendorp was the lowest. Similarly, the highest mean income in Mashishimale was in the more established R1, while households in Mophela proper, which had proportionately more male heads than Sankontshe, reported a higher mean income.

The per capita income, taking household size into consideration, provides a simple comparative measure of the material circumstances in these communities. Based on reported survey data Ebenhaeser had a per capita monthly income of R743, Mashishimale R451 and Mophela R325 in 2008. To put the reported per capita monthly income into perspective, the minimum cost covering essential items adjusted to 2006 prices by StatsSA yielded a poverty line of R431 (StatsSA & National Treasury, 2007). In at least two of these communities the situation is precarious.

Using a more differentiated model based on 2008 currency levels, the Presidency development indicators for 2009 show that 22% of the country's population live on a per capita monthly income of less than R283 (20th percentile) (RSA, 2009:26). The proportion of the population living below a per capita monthly income of R388 (40th percentile) rises to 39%, and those below R524 (60th percentile) to almost half (49%). The per capita monthly income calculated from survey data for Mophela places it below the R388 threshold, while that of Mashishimale is below the R524 threshold. It is evident that the level of income is low and that a significant number of households face economic hardship that makes everyday survival immensely difficult.

Eliciting household income is a challenging endeavour as it infringes on privacy and is consequently a sensitive matter, often raising suspicion. Investigating patterns of expenditure provides additional information on a community's material conditions. Triangulating income and expenditure smoothes out distortions that may occur in a context of non-disclosure, particularly where the impoverished may feel that full disclosure may lead to the loss of some of the benefits they receive.

6.2 Average household expenditure

Although it is unlikely that a one-off survey asking interviewees about their expenditure in the preceding month can produce an exhaustive account of household expenditure, it does give a reasonable indication of the resources at the disposal of households.

Food is the largest component of household expenditure. Clothing, another essential item, constitutes about 10% of household expenditure. Together these items accounted for almost half of household expenditure in Ebenhaeser and Mashishimale and about a third in Mophela. Essential services such as transport (enabling access to work), education and municipal services (water and electricity) constitute slightly more than a quarter of household expenditure. Transport and education constitute significant items. The amount spent on transport reflects the distance of the community from nearby towns which have a better infrastructure and potentially better job opportunities. Expenditure on municipal services is low, in part due to limited resources which, for instance, curtail the usage of prepaid electricity. Households in Mashishimale and Mophela have invested in education much more than in Ebenhaeser, as a livelihood strategy, with an expectation that better employment would yield dividends. However, the current difficult labour market conditions lead to immense disappointment in this regard. In discussions with youth during fieldwork, many were keen to access training they saw as potentially strengthening their formal employment opportunities. They expressed less interest in agricultural training compared to other options. These preferences suggest aspiration towards professional and white-collar employment in the first instance. Generally, households have limited discretionary expenditure and limited capacity to save. It can be assumed that inflationary pressure on the basic commodities consumed by households in the community will exacerbate pressure on expenditure, with regard to food and energy (including fuel) in particular.

The mean monthly household expenditure reported was R1856 in Ebenhaeser, R1585 in Mashishimale and R3308 in Mophela (see Table 5). This translated into a monthly per capita expenditure of R599 in Ebenhaeser, R428 in Mashishimale and R735 in Mophela. In two instances monthly expenditure recorded in the survey was above a monthly per capita of R524 based on 2008 currency levels (RSA, 2009:26). Although the per capita expenditure in Mashishimale fell below the R524 poverty line, its income was slightly above this measure.

Reported monthly expenditure in Mophela exceeded reported monthly income by a considerable margin. Its per capita monthly household expenditure, taking household size into consideration, remained the highest for the three communities, but narrowed the gap between them somewhat. Given the location of Mophela, within the boundary of a metropolitan municipality, albeit on its periphery, a higher per capita income (not recorded) as well as expenditure was to be expected.

Approximately two thirds of households in Mophela reported that their expenditure for the preceding month had been higher than was normally the case. Households in this community service more debt proportionately than those in other communities, suggesting they may use loans as bridging finance. However, with seven in ten households (72%) (not shown here) claiming that their income in the preceding month was comparable to other months, it is unlikely that loans could finance the gap between reported income and expenditure. This suggests that household income was concealed or under-reported in this community.

A clear trend was evident with regard to additional sources of income for male- and female-headed households (see Table 6).

Besides wages and self-employment as sources of income, social grants provide a key cash anchor and lifeline to households in these three communities, in a context where there has been a significant decline in formal employment opportunities. According to the Presidency development indicators (RSA, 2009), in 2009 such social welfare assistance constituted 5.5% of South Africa's GDP. In all three communities the proportion of households that had at least one member receiving a social grant was larger than the proportion that had at least one adult member in formal employment.(see Table 7) The importance of social grants for preventing destitution and providing households with some relief cannot be overemphasised. Many children receive a child support grant, and a large number of state pensions are awarded as well (Bryceson, 2004; Seekings, 2005). Ebenhaeser, with an aged dependency ratio double that of Mashishimale (which was the same as recorded for the country in the 2001 Census), and more than double that of Mophela, receives proportionately more pensions than the other two communities (see Table 1). During fieldwork it was observed that ‘all-pay day’ – an informal term in Ebenhaeser for the day social grants are paid out – is keenly anticipated and engenders a festive mood in the community.

Table 7: Employment by sex of household head

	Ebenhaeser		Mashishimale		Mophela
	Male	Female	Male	Female	Male	Female
Employed	55	44	72	53	69	62
Seeking work	66	57	82	75	69	44
Unemployed	72	66	76	75	74	48
Entrepreneurs	13	6	22	19	16	10

A comparison of household employment reveals a distinct pattern. Proportionately more male-headed than female-headed households have at least one adult member who is employed. However, these male-headed households also have proportionately more adult members who are unemployed. Male-headed households are generally larger than female-headed households (see Table 7).

7. Income sources

Data on income and expenditure in these communities reflect the key role played by cash. Access to formal employment and a permanent source of income is a key livelihood strategy for households. The proportion of households in these communities with at least one adult member formally employed was 51% in Ebenhaeser, 65% in Mashishimale and 67% in Mophela. A smaller proportion of households in Ebenhaeser than in the other two communities derived some of their income from an adult in formal employment. Four factors probably explain this situation: Ebenhaeser has a lower mean household size, more aged inhabitants proportionately, is further away from larger towns where employment opportunities may be sought, and has limited access to public transport. In Mashishimale and Mophela approximately a third of the households did not have access to income derived from formal employment. The corresponding proportion of households in Ebenhaeser was almost half.

About two thirds or more of the households in these communities – 63% in Ebenhaeser, 73% in Mophela and 76% in Mashishimale – had one or more adult member seeking employment (see Table 7). This emphasises the importance attached to employment as a source of cash income. Hajdu (2005) observes that job seeking appears to be more cost effective than farming for individuals in these households. A drawback of farming is that it requires capital for inputs and unpaid labour from households.

The few available employment opportunities in the communities are linked to education and health care, and to some positions in local governance structures that are remunerated. While local business enterprises also provide some employment, most of the job seekers commute to places of work within the municipal boundaries or to adjacent municipalities. In all three communities, unskilled and typically casual jobs constituted the largest single category of employment, encompassing more than half the employees, working as labourers. A number of inhabitants work on farms or as domestic labourers in town. The rate of unemployment for youth remains high, emphasising their vulnerability.

The informal sector is an alternative source of income, mostly organised around selling and to a lesser degree around providing services locally. Generally, fewer households were active in this sector than in formal employment. However, this does not detract from the crucial role this sector plays in redistributing income in the community, particularly in instances where households have limited access to other sources of income. Mashishimale had the highest number of households, proportionately speaking, that derived some of their cash income from such activities (21%). Mophela, by contrast, had the lowest proportion of households supplementing income in this way (4%).

8. Conclusion

Households do not use subsistence farming as a key survival strategy, but only as a supplementary means of livelihood, in spite of the precarious material conditions many face. This tendency is out of necessity rather than choice. Agricultural produce does play an important role in alleviating food shortages and mitigating difficult material circumstances, but in most instances does not yield a surplus. Most critically, where households have access to fields they lack capital, a regular water supply, labour and fencing. The risks associated with investing in farming in part explain why securing employment rather than engaging in small-scale agricultural production remains the most important livelihood strategy for these communities.

Young people especially are not familiar with a way of living that involves reliance on agricultural activities. Instead they aspire to employment. The perceived promise of employment is an incentive for many to invest substantially in schooling, in spite of a very low return on such investment. Attempts to secure formal employment within the municipal district cap all other strategies of livelihood in the three communities surveyed. Indeed, households that can access a regular income through permanent employment have a much better chance of meeting their needs and accumulating both financial capital and skills that can be transferred to later generations.

Atkinson cites Minister of Rural Development and Agriculture Gugile Nkwinti's assessment of restitution: ‘Food security and economic growth are being undermined by the collapse of more than 90 per cent of the farms that the government [has] bought for restitution or redistribution to victims of apartheid’ (2010:364). This reflects a shifting emphasis towards viewing land primarily as an asset, requiring development. Land is increasingly viewed as having commercial value. There is a strong incentive from the state's side to promote commercial agricultural farming, whether through large-scale concerns or by supporting small-scale emerging commercial farmers.

Yet, as our evidence shows, most people in rural communities are not oriented primarily to agricultural production. In spite of an insistence by communities that land should be ‘returned’ to its rightful ‘owners’, ordinary people remain alienated from land in a changing economic and political context. This puts decision making about how to use restored land in the hands of a few individuals managing commercial agriculture, who are seen as acting on behalf of everybody else. Such use of land is therefore a policy incentive that proves the viability of land restitution from the perspective of the state. This policy considers vulnerable communities, who are not able to use the restored land fully, to be a constraining factor and puts pressure on them to concede to its use for commercial agriculture.

What is happening is not simply a matter of ‘those in power making decisions for the poor’, or an underestimation of ‘the peasant question’, as some suggest. Although commercial agriculture is a key imperative of land restitution for political reasons from the state's perspective, it also is an ‘option’ imposed by the pressure of circumstances at the level of ordinary people's livelihoods. The real difficulty remains one of whether the dividends yielded by these commercial activities will be adequate to meet these community's expectations. Should the restored land be divided equally, the allotment size would range from 1.96 hectares in Ebenhaeser to a high of 8.68 hectares in Mashishimale (the most favourable calculation) per household. Such allotments fall short of the size necessary to ensure subsistence in the Eastern Cape, as calculated by Tomlinson (1990). Given the natural population increase in these communities and the size of the restored land, it is unlikely that the high expectations pinned to land restitution would be met. Kepe and Cousins (2002), rightly argue that extensive redistribution of land is required to address poverty in these communities and ensure sustainable livelihoods based on agricultural production, should that be the development route opted for. Walker's note that ‘the connection between land rights and enhanced livelihoods or economic growth tends to be assumed rather than examined’ (2005:805) should also be kept in mind.

Addressing the paradox of political inclusion and economic marginalisation of historically dispossessed rural communities is a key challenge facing the post-apartheid South African democratic state. While the moral and political imperatives for land reform remain undeniable, these have to be balanced against the need to cultivate agricultural land productively and the wishes of the claimants concerned.

Acknowledgements

This study is based on a collaborative research project by the Department of Sociology at the University of Pretoria. Dr Bomela, a member of this research team, has subsequently joined the Nelson Mandela Metropolitan University, Port Elizabeth, as a Research Manager. The authors would like to thank postgraduate students from the Department of Sociology for assistance in monitoring fieldwork during the surveys, coding interviewee responses and capturing data, and the Canadian International Development Agency (CIDA) for funding the pilot project.

Notes

¹Afrikaans HOP, Heropbou en Ontwikkelingsprogram.

²The South African Development Community (SADC) Centre for Land-Related Regional and Development Law and Policy (CLRDP) at the University of Pretoria and Commission for the Restitution of Land Rights selected the case studies (CLRDP, 2010a).