Copyright Sociological Research Online, 2000


Charles Crothers (2000) 'The Effect of Community Context in The South African Land Reform Programme'
Sociological Research Online, vol. 5, no. 2, <>

To cite articles published in Sociological Research Online, please reference the above information and include paragraph numbers if necessary

Received: 8/8/2000      Accepted: 5/9/2000      Published: 6/9/2000


Many colonised peoples have been concerned about the return of their land. The context in which their land is returned is undoubtedly important in shaping whether or not land reform is successful. In turn, successful return of land may benefit by guidance from social science studies. In this paper I attempt to add to the limited literature which seeks to understand the factors which shape the success or failure of land reform projects.

Zimbabwe has had some historical success with land settlement: see Cheater, 1984. But, the recent violent, vigorous and highly politicised land reform campaign in Zimbabwe has overshadowed the rather more peaceful South African attempts at land reform. It is clear that the Zimbabwean process has been fraught with politicisation and corruption, and it is doubtful how serious this attempt at providing institutionally sound land reform this process has been. Whatever its local effects, the spectre of the Zimbabwean conflict has cast a lurking shadow over the South African farming and land reform situation. Nevertheless, the relative calm of the South African situation has allowed its land reform process to develop more quietly, and it is possible to better gauge what has worked and what has failed to assist specific land reform projects. Further, a research programme has been developed in South Africa which should be able to assist in this process.

As with many other situations where there is a strong need for land reform, the seeds of the South African situation were sown over many generations of gradual dispossession of areas of land from their indigenous owners. This sorry history has been recorded in many places. Suffice, then, to provide an overview summary: "The history of the land in South Africa has been one of massive state intervention. The land was taken from the initial inhabitants of the country by force and trickery, not through 'fair' contractual relations. And then as African farmers struggled (often amazingly successfully) to re-establish themselves within the new legal restrictions imposed on them as blacks, so political intervention by a white oligarchy was used again and again to prohibit their right to enter into ordinary contractual relations. The tenure system of private property is built on racial dispossession and racial exclusion; the primacy of race over contractual considerations has been asserted by the state in a blatant and unashamed way for centuries. Since the Land Act of 1913 which did reserve some land for blacks, the remaining black farmers outside these areas were vulnerable to sudden and irrevocable dispossession" (Claasens in De Klerk, 1991:50).

Land reform is situated in a complex relationship between existing farming and demands for alternative ownership. Some of the demand for land emanates from those already involved in working on the land, and some from non-owners. The reaction of existing land-owners is in part dependent on the relations they have with their workers, and perhaps also with potential occupiers.

It is important to establish some of the characteristics of the current situation of farming in South Africa as an important context within which land reform must take place. Although commercial farming in South Africa has been seen as a 'white' activity, white farming has nearly always been accompanied by large and black (in some areas, coloured or Indian) farm workforces which have often been harshly treated. One measure of the difficult situation which has ensued has been the number of farmers killed over the last decade. The onslaught of farm attacks has in part been tactical, since individual white families living in the middle of their farm are perhaps softer targets than were they living in an urban setting. In part, too, attacks on farmers have been politically motivated and ideologically/symbolically defined: 'one settler one bullet' was an expression aimed straight at the heart of (illicit) white occupancy of land. However, it is also an indicator of difficulties faced by farm workforces, and their attendant issues of access to side-land for home cultivation, housing etc. The spate of farm attacks must be set alongside the often-reported (but not statistically cumulated) reports of bad treatment of farm workers, and the frequent dumping on the roadside of farm workers by farmers intent on ensuring that their tenants's occupancy rights do not become legally entrenched.

Attacks on farms and smallholdings continued through the nineties at a high rate. The South African Police Service (SAPS) record these separately (SAPS: 1999) and sum up 769 attacks during 1998, with 142 murders resulting. Attacks occur across all provinces. For the first quarters only, farm attacks in 1998 were 53% up on those for the same period in 1997 and the 1999 figures pushed the proportion increase up another 38%.

Another indicator that all is not well down on the farm are recent newspaper reports of farm occupancies in South Africa. This is not only a copy-cat of Zimbabwean tactics, as farm invasions go back through the decade of the nineties. There are widespread newspaper accounts of squatting on land, with estimates up to 10,000 in this situation. Farmers is some areas report destruction of crops and infrastructure, and a level of harassment which undermines their farming operations. Much of this squatting involves people moving back to land from which they had been disposed in the last few decades, impatient with the slowness of restitution processes. In some cases, though, unscrupulous leaders illegally encourage squatting and 'sell' land which is not in their control.

Land reform initiatives add a further aspect to this volatile mix. The remainder of this paper will focus more directly on these.

The Community Context of Poverty

I make the assumption that land reform schemes can be usefully appraised within the context of other poverty alleviation strategies. Most studies of poverty examine the relationship between assets and income (and also the generation of substitute income) at the household level. In addition, especially as part of the drive to understand the gendered nature of poverty, there has grown, more recently, a strong interest in the intra-household distribution of work and of rewards. However, I pursue neither of these two more well-worn paths. Rather, at the other end of the scale, I am endeavouring to also examine supra-household, community-level, influences on poverty. The community context can play a major role in the development processes of the households within communities. In some part, this community level is the aggregate effect of the characteristics of the households which is important, in some part it is the suite of institutional supports the community can collectively call upon that is important, and in some cases it is their experience of interrelating with each other: the social capital inherent in their relationships (see Crothers, 1996).

Possible effects of each of these three types of mechanism might include:

But, as yet, too few studies at this level have been carried out for social scientists to be clear about how useful this level of analysis is in providing an understanding of poverty. But this level of analysis is particularly pertinent in relation to land settlement schemes, which tend to be settled in a collective form. So, in terms of my broader research program into community-level effects on poverty, examining land reform groupings is a particular pertinent research-site.

The Institutional Framework of South African Land Reform

Policy is built on the assumption that there clearly is a considerable demand for land in South Africa. A study carried out by Marcus et al (1997: 16) has in fact shown that some 68% of black rural households desired farmland. On average the typical household desired a small block of land (1 ha. or less), although some were keen on obtaining large blocks of land. Marcus et al report a considerable regional variation, which they attribute to the varying likelihood amongst provinces as to the viability of farming. They also report that the quantitative estimate of demand could be subject to misinterpretation as their focus group data revealed that the demand for land was quite universal.

To meet at least some of this demand a three-pronged approach has developed (May, 2000: 242):

As a whole, the land reform programme is designed to address the injustices of the past, provide for a more equitable distribution of land ownership, reduce poverty and contribute to economic growth, ensure security of tenure for all, support sustainable land-use attains and facilitate rapid land release for development. (May, 2000: 242)

Estimates of numbers of people who might potentially benefit from these programs (over the next ten years) are:

The ability to deliver these programmes depends on the capacity of the state, the capacity of provincial and local governments, budget funding and availability (in at least the first instance) of state land (not all of which is suitable for redistribution). In the period since the advent of the Government of National Unity (GNU) the capacity of the state has steadily built up, although not entirely matched by the build-up in capabilities of its provincial and local government partners. A minuscule, but steadily increasing budget, has progressively year by year been able to implement rather more-on-the-ground delivery of reform, as year by year statistics show. However, progress remains slow.

Nevertheless, there is little use resettling people if the projects are doomed to failure once the beneficiaries are settled. Complementary measures are very important: including ".. extension support, improvement of marketing infrastructure, access to credit and other financial services, and the creation of off-farm income opportunities" (May, 2000: 243). Further "Post-settlement support, in terms of training, finance and advice, will be crucial" (May, 2000: 244).

Internationally, tenure reform programmes are renowned for their limited success. Too often, policies have worked only or mainly to the benefit of the rural elite or the local bureaucrats. While it is clear that the South African programmes are concernedly designed to not fall into such traps, the costs are high of dealing with the myriad of complexities that must be tackled to ensure a high likelihood of successful projects. One of the down-sides of trying to carry out land reform well is slow progress.


To assess the success of the land redistribution programme in South Africa a third assessment study was launched in 1999. The data for the Monitoring and Evaluation (M&E) study was undertaken by CASE (an independent research company appointed through standard government tender process) with fieldwork being carried out between September and November 1999.

In addition to a household questionnaire (which includes a considerable level of information about each person within the household), a community questionnaire is central to the M&E process. This covers questions on food prices, communal projects, communal agricultural activities, project income and expenditure, loans/ grants/ subsidies and community organisation. The unit of analysis in the community questionnaire varies, with the focus shifting variously between joint production units, community property associations and the overall institutional framework of the community. Where possible, the questionnaires have been designed for comparability with other South African surveys, such as the October Household Survey. The community information was collected from informed respondents from the land administration association, project managers, important members of the beneficiary community and service providers.

The research design is complex and ongoing with community and household questionnaires administered to samples of existing and new projects, with the community questionnaire re-administered every 3 years in previously-sampled projects. Unfortunately, it was found that for the 1999 data round the community questionnaires had a high incidence of missing data. This was unexpected since the experts being interviewed might have been expected to have returned the most reliable data amongst the various data-sources being tapped. This appears to have been a problem with the quality of the field-work.


The main report of the 1998/99 Evaluation concludes as follows. "This study has shown that there has been an improvement in both the performance and impact of the land reform programme since the previous monitoring activities were undertaken. The rate of delivery has increased, targeting of the most poor has taken place, both agricultural and non-agricultural production is occurring, service delivery to land reform beneficiaries seems better than to the rural population as a whole, and there is less evidence of institutional problems. However, poverty levels among land reform beneficiaries remain high, as do the levels of dissatisfaction that they express. Many projects do not yet show any signs of economic potential, and many participants in the land reform projects have little knowledge of the management of the project and how funds have been utilised. This opens opportunities for corruption and the misuse of community funds" (May et al, 2000: 160).

The May et al (2000) report contains data from the population of projects, which is further analysed in Tables 1-3. In Table 1 the provincial distribution of number of projects and number of beneficiaries is shown. In Table 2 this is related to the share each province has of people, poor people and non-urbanised people. The proportion of rural people and the proportion of poor people are highly correlated at the provincial level, although while share of rural population is moderately correlated with share of land projects, the correlation with share of poor people is small. This suggests that the distribution of projects is not well related to any of these other provincial characteristics and therefore it is not especially well targeted: at least in terms of the provincial level of aggregation. However, this spatial level is probably too coarse for accurate results to be obtained.

Table 1: Provincial Patterns: Shares of National Totals
Table 2: Provincial Shares: Correlation Coefficients
Table 3: No. of Beneficiary Households

May et al report a household level regression against whether or not the project was successful in terms of household characteristics. The measure of success was whether or not the project had made a profit. A similar exercise is now carried out, but at the community/ project level and using only communal/ project characteristics as predictors. Interestingly, this dependent variable reveals an even split between those communities which made a profit and those which didn't. (In the tables, profitable communities/ projects are scored 1, and non-profitable ones are scored 0.) Instead of multiple regression analysis the rather more robust data-analysis procedure of multiple classification analysis (MCA) is deployed.

Where complex sets of categories have been identified, these have also been recoded into rather more simple alternatives, usually where the main category has been counter-posed with all the other mall categories lumped together. A commentary on the results of Table 4 now follows:

Table 4: Community Characteristics

The type of legal entity most preferred is the trust, except in Guateng and Northern Cape where the Common Property Association (COA) is more preferred.

Just over half the communities did not gain any profit, with the more successful projects split evenly between those reinvesting and those paying out. Few obtained cash loans, and very few other forms of government grants. Most had funds remaining from their Land grant. Meetings were usually frequent, with over one half the communities having met in the previous month. Two-thirds claimed regular feedback meetings. Most had met before making their claim. While on the whole there are regular meetings of communities and committees, regular report-backs from management committees were less frequent.

The legal entity or legal committee was seen in most communities as the most important mechanism for solving disputes, for being in charge of the land and for being in charge of water. In three-quarters of the cases, the legal entity was seen as working harmoniously. It is interesting that the legal entity is so-widespreadly seen as important, as opposed to traditional leaders or other authorities which might be supposed as holding sway in such rural areas. This suggests that projects often involve 'new' political and institutional structures and/or that they are located in newer locations. Land projects are not often under the sway of traditional structures. Of course, it may well be, however, that in practise traditional leaders occupy important positions within these new structures. Despite the considerable role accorded them, in most community situations the legal entity did not hold a monopoly of community power, as other important groups and committees could be identified. In most communities it was felt that conflict could be resolved using clear rules, and that in practise those acting badly could be kept in check, although in rather fewer cases was it felt that there was a high level of trust in the community.

Less than one third of respondents were aware of the way in which the Land Grant funds has been used or even what type of land management structure has been adopted. (This is important since May et al had established that an important determinant of project success - amongst household predictors - was knowledge of the rules governing the project.)

The majority of land reform beneficiaries reported that the past year has been peaceful. Most reported at least moderate levels of trust in the community (except in the North West, Northern and Northern Cape provinces). The conflict arising related to conflicts amongst trustees or committee members; the distribution of products or income from projects; and the allocation and use of land.

Table 5: Factor Analysis of Community Characteristics

A factor analysis was carried out in order to endeavour to group clusters of variables and reduce redundancy (Table 5). The importance of each factor identified through this procedure cannot of course be 'read off' the table since the rank of each factor depends on the number of related (even redundant) questions that happen to be asked in the questionnaire. The first factor relates together various aspects of the functioning of the community and project: regularity of meetings and reporting, authority, procedures and trust. The second factor relates to funding: whether funds remain, especially from the land grant and whether or not other funding has been obtained. The third factor emphasises size, but also whether other income had been obtained. A fourth dimension concerns the various questions which deal with the relationship between the more important groups in the community, especially the role of the legal entity. The last clear factor merely identified proportion of women involved in the legal entity as a measure not directly relating to any other in that set, and therefore not apparently strongly related to other aspects of community and project funding (or at least those measured in this questionnaire).

Table 6: Analyses of Variances of Profit in terms of Independent Variables

The next step in the analysis is to ascertain which of these variables are associated with profit-making (Table 6: note that means closer to 1 indicate financial success, and means closer to 0 indicate financial failure). The results of the factor analysis are used to guide the data-analysis, and to ensure that there is not too much redundancy in the interpretation of the data. Each of the variables in the first cluster has some relationship with profitability, although none are strong, and some work in surprising directions. Whether or not it is the legal entity which is dominant in the community proves to have an entirely random relationship with success, so that several of the variables relating to legal entities have no net effect at all.

Some results are interestingly counter-intuitive. For example, non-harmonious entities are more likely to be successful! Another surprising finding is that level of reported trust in the community seems to have no effect on profitability.

Those communities/projects able to attract outside funding do better. So do those who have used up their land grant money: perhaps an indicator of greater length of operation and therefore a better chance of success.

Size proves to be a strong predictor. It may also be quite complex as it is mainly the largest or the smallest size categories which reveal the highest degree of success. Proportion of women involved with each project is interesting but not definitive: there is apparently more success the higher the proportion of women, but this relationship is not statistically significant.


The social characteristics of communities do seem to have an effect on the performance of economic activity within their boundaries. Even with a relatively small sample of communities, this study has been able to show that community effects are quite powerful predictors of whether or not a project is financially successful. However, some of the community predictors identified as being more tightly tied to success, were surprising, and some of the postulated effects rehearsed in the literature seemed not to have the predicted type of effect in this data set.

Project size was important, with number of participants in the project, in particular, proving a strong predictor. Institutional structure seemed to be considerably less important, with institutional structures perhaps proving to be robust, despite different ways in which social processes occur within them. Shared community capital, such as level of trust, was decidedly unimportant. In fact, some degree of disharmony and tension seemed more likely to be associated with success, rather than lack of success. More analysis is needed to entirely tease out these linkages.

Using land reform projects as a research-site for studying community effects has been a useful addition to my overall research programme and to the literature in this area. Besides replication across a broader range of conditions and utilising larger sample sizes, further exploration of social and community factors and their effects are required. In particular, future analysis should attempt to more effectively separate out the different effects (and interaction between) of social factors and of economic factors on project success. It is difficult to assemble a useful story about the relationship between social factors and project success without also including an investigation of the effect of economic factors and of course of household characteristics. It is not yet clear how important social factors are in this broader story. It may be that social factors are more or less important than economic, and it is likely that both are needed in predicting success. However, the relationship may be more complex. The literature on social capital may afford a clue to the likely complexity. Social capital may well, as Bourdieu has suggested, act in particular as a 'multiplier', adding to the economic, but requiring some economic base on which to multiply. But these questions await further investigation.

Land is important in many people's lives: it is a tragedy if their hopes are dashed by being settled in projects which do not then prove capable of supporting them. Part of the delicate art of developing successful projects is to ensure that their community context is supportive. But, it seems from the analysis carried out here that much more social knowledge is needed on how to achieve this.


CHEATER, Angela (1984) Idioms of accumulation : rural development and class formation among freeholders in Zimbabwe Gweru : Mambo Press.

CROTHERS, Charles (1996) "South African Poverty in its Community Context" Paper for LAPC Conference on Rural Livelihoods, mid-Rand.

De KLERK, Michael (ed.)(1991) A Harvest of Discontent: the land question in South Africa Cape Town: Idasa

MARCUS, Tessa et al (1996) Down to Earth: land demand in the New South Africa Durban, Indicator Press

MAY, J et al (2000) Poverty and Inequality in South Africa: meeting the challenge Cape Town: David Philip.

SOUTH AFRICAN POLICE SERVICE (1999) Attacks on Farms and Smallholdings: Report of 1999. <>.

Copyright Sociological Research Online, 2000