The Constitution of 1996 advocates a radical break from the past racial distributional and ownership patterns of land. Section 25, known as the 'property clause', undergirds all the land reform legislation and programmes that followed. It begins with protective property clauses and goes on to bind the state to initiate a comprehensive land reform programme, distinguishing between redistribution, restitution and tenure reform:

The first clause has proved controversial. Some political groupings on the left criticise this clause for protecting white, registered rights as part of the negotiated political settlement to end apartheid, and hence call for section 25 of the Constitution to be amended to allow for expropriation without compensation [3]. An alternative interpretation by some of the country's top constitutional property law scholars would suggest that this clause should not be abandoned as it also protects new interests in property among black rights holders arising from the land reform programme [4]; while more recently some land rights activists point out that this clause could be a powerful vehicle to strengthen unregistered property rights, since the Constitution does not state that only registered property is 'property'.  Although the state has consistently failed to see the potential of unregistered land rights qualifying as 'property', this line of thinking is informing alternative approaches to land tenure reform [5].

As is evident in sub-clauses 2-5, property may be expropriated for 'a public purpose' or in the public interest, which includes land reform, but is subject to various conditions including compensation. Section 26 provides for rights to adequate housing and prohibits arbitrary evictions.

Parliament enacted a draft of laws to provide for:

(1) Redistribution and restitution, viz.

• Restitution of Land Rights Act 22 of 1994
• Restitution of Land Rights Amendment Act cc of 2014
• Provision of Land and Assistance Act 126 of 1993, which is the operating law for redistribution.

 (2) Provision of base-line protection against eviction or arbitrary dispossession applicable to particular categories of people:

• Farm workers, farm dwellers and labour tenants on privately owned land;
• Residents in coloured rural areas, mostly former mission stations;
• People living under customary tenures ('communal tenure') in the former Bantustans;
• Occupants of (rapidly expanding) informal settlements in urban areas.

The relevant tenure laws are:

• Land Reform (Labour Tenants) Act (LTA) 3 of 1996
• Interim Protection of Informal Land Rights Act (IPILRA) 31 of 1996
• Extension of Security of Tenure Act (ESTA) 62 of 1997
• Prevention of Illegal Eviction from and Unlawful Occupation of Land Act (PIE) 19 of 1998

Successful land restitution claimants and beneficiaries of land redistribution could own land collectively in terms of the:

• Communal Property Associations (CPA) Act 28 of 1996

Older laws passed toward the end of apartheid allow for upgrading to, or maintenance of, titles:

• Upgrading of Land Rights Act 112 of 1991 (ULTRA)
• Land Titles Adjustment Act 111 of 1993 (LTAA).

There remain significant barriers to legal recognition of off-register rights. Protected unregistered rights are relatively secure at local level, but not so against third parties or the world at large, e.g. private or state investors. Upgrading these rights by registration to title is not feasible unless major adaptations were to be made to the property system. Titling is the policy ardently pursued by market economists and the state. Apart from the expense and complicated procedures of registration, what titling advocates ignore are the significant mismatches between 'living law' practices (modern 'customary' norms in rural and urban contexts) and the system of title deeds. De facto rights are held in extended families and often involve multiple and overlapping use rights, for which there is no recognised holding mechanism in law [11].

In terms of the Deeds Registries Act, a land may only be registered in ownership when certain stringent conditions have been met, including the survey of the land parcel in terms of the Land Survey Act. The implications are that for rights to be registerable, a one-to-one relationship between a surveyed property object and a legally identifiable juristic property holder, single or corporate must be established. These 'linear' property relationships do not work in situations where customary property relationships described above, where property has less clear-cut spatial and social boundaries, with layered access and use rights held across inter-generational families. These relationships tend to persist even after the title deeds have been issued, which means that the titles soon bear little resemblance to the arrangements on the ground [12].

What is more, in cases where rights are completely subservient to the dominant registered title of the owners, such as in the case of farmdwellers, there are virtually no means of strengthening those without converting them into registerable rights of ownership. This is simply not feasible for the most part. Millions of farm dwellers are thus extremely vulnerable to the overriding rights of the registered owners, and anecdotal evidence suggests that the legislation ESTA, that protects rights of farmdwellers, may have exacerbated both illegal and legal evictions.

Communal tenure reform has been deeply contentious and prolonged. The Communal Land Rights Act 11 of 2004 (CLRA) was declared unconstitutional. Opponents of the law considered this judgement a victory, since it the law was seen as a re-enactment of apartheid systems of tribal authority rule over restricted areas of the country. They argued that the CLRA gave traditional councils (known as 'tribal authorities' under apartheid) wide-ranging powers, including control over the occupation, use and administration of communal land [13]. The law was not, however, struck down on substantial grounds, but rather because the Court found that the correct procedures for its enactment had not been followed [14]. 

Indeed, after several attempts, a new Communal Land Tenure Bill has recently been released for comment. It proposes a hybrid form of registered communal titling for communities that acquire a juristic personality, and who may then be issued with a 'Deed of Communal Land'. Individual breakaway titling is legally possible but practically unfeasible. The Bill proposes a plethora of new management structures, which though theoretically allowing for some choice in governance structures, will almost certainly empower existing traditional authorities, since the Bill:

• fits into a broader framework of traditional governance in the form of the Traditional Leadership Governance Framework Act of 41 or 2003 as amended in 2009 (TLGFA) and the Traditional Courts Bill of 2017 (TLB) [15].
• recognises traditional authorities (renamed Traditional Councils in terms of the TLGFA) which democratically-minded citizens regard as apartheid-created political structures that have been given broad-ranging powers over people living in the former homelands, including land administration functions, without sufficient means to opt out of the system;
• potentially endows Traditional Councils with powers of collective ownership over land at the expense of the rights of individual family units.

Critics are concerned that Bill will result in concretising apartheid boundaries and structures that were based on territorial and ethnic geo-spatial entities, with highly unequal spatial and political power [16]. The Bill potentially plays into identity politics and gives people insufficient means to opt out of the imposed system. Critics point out that, taken together, the three pieces of legislation do not provide sufficient means to hold traditional structures to account or counterbalance their power over local governance and justice by providing recourse to other state institutions nor ensure adherence to constitutional principles [17].

Some land rights advocates recommend an entirely different approach from laws of specific application that apply only to geo-spatially designated areas that hark back to colonial and apartheid institutions and boundaries. A key proposal is for:

• Strengthening land rights by tightening up existing tenure laws.
• Adopting a 'Land Records' system to provide legal-administrative support to ensure that these rights are legally recognised without the immediate requirement of registration through processes of country-wide enumeration, ultimately to be adjudicated according to new forms of evidence. The system should be capable of recording family holdings appropriately designed for all currently off-register rights across the board, i.e. not confined to  'communal' areas. The aim would be eventual integration of land administration systems in the country [18].

1. Redistribution

A recent review of current redistribution policies and practices has charted the drastic swings in land reform policies that have accompanied each change in presidency since the transition to democratic governance in 1994 [37]. The early post-apartheid model of land redistribution adopted by the ANC government drew from the influence of the World Bank, which advanced state-assisted land purchase and transfer of title to beneficiaries. This took the form of small grants to poor households to buy modest areas of land for settlement and small-scale farming. From 2000 the policy shifted towards promoting black capitalist farmers, and providing larger land purchase subsidies to those with their own means to engage in commercial production. From 2011 the state has adopted the willing buyer, willing seller approach. Now the state has itself become the purchaser of land, acquiring land for redistribution to beneficiaries without transfer of title, thus moving away from the original private ownership model to state leasehold, with very little substantive adherence to the original goals of land reform [38].

The policy is guided by a Proactive Land Acquisition Strategy (PLAS), which empowers state officials to buy farms on the open market and allocate them to selected beneficiaries. In 2013 the emphasis shifted to 30-year leases renewable for a further 20 years before the state will consider transferring ownership to them under the State Land Lease and Disposal Policy (SLLDP). To qualify for on-farm infrastructure and production support, under a Recapitalisation and Development Programme ('RECAP', see below), ‘beneficiaries’ are required to enter into a partnership with a ‘strategic partner’ – i.e., a farming or agribusiness company – in a mentorship or joint venture arrangement.

Hall and Kepe conclude that "[w]hile discursively framed as part of a radicalisation of the reform process, the redistribution process appears to be narrowing and is showing signs of elite capture"[39]. They reveal that the policy is applied highly unevenly and unfairly, and in most cases the so-called beneficiaries have no leases at all, but occupy the land as caretakers under highly insecure tenure with no records. In most cases the only meaningful contractual arrangements are between the state and the 'strategic partners' or 'mentors'.

Land rights advocates maintain that a key constraint on effective land reform is the absence of a law on land redistribution to give full effect to section 25(5) of the Constitution, recommending a framework law to address the gaps in relation to: guiding principles; definitions of key terms such as ‘equitable access’; clear institutional arrangements (particularly at district level); requirements for transparency, reporting and accountability; and other measures that promote good governance of the land reform process as a whole, as well as land redistribution in particular, and to bring all the existing anchor laws into line with the framework law, thus ensuring overall consistency and coherence within the land reform programme [40].

2. Rural state subsidies

It is widely acknowledged that land reform has not unlocked the potential for dynamic economic growth and employment creation, especially in the agricultural sector. The Recapitalisation and Development Policy Programme (RADP or ‘RECAP’) of 2009/2010 replaced all previous land reform funding programmes. Funding is for a maximum of five years and beneficiaries must have business partners, who assist with 'business plans', recruited from the private sector, as mentors, or within share-equity schemes, or through contract farming. Evaluations of RECAP, particularly one conducted by the government Department of Planning, Monitoring and Evaluation (DPME) [41] have found the programme to be wanting in many respects and suggested it be scrapped. Critics at PLAAS maintain that RECAP is an attempt at ‘commercialisation’ of land reform projects, where large sums are spent on relatively few beneficiaries, few jobs have been created, and access to markets remains limited [42]

The programme has since been moved away from the Department of Rural Development and Land Reform to the Department of Agriculture, Forestry and Fisheries, though the latter complained it was an 'unfunded mandate' as the transfer came without funds [43].

State subsidised rural programmes focus on agriculture, but research suggests the need to shift state investment to the non-farm sector given the poor returns. The Comprehensive Agricultural Support Programme (CASP) — that applies only the communal areas — has been running since 2004, providing grants to applicants, so far having allocated R750 million to the programme.

Many evaluations of state programmes, including one published by Presidency [44], conclude that despite CASP and RECAP, there has been a lack of transformation in the commercial agricultural sector, and "a large number of land reform beneficiaries have not been able to use the land productively, partly due to inadequate infrastructure, inputs and technical support after finalising settlements”[45]. Prof Ben Cousins of the Institute of Poverty Land and Agrarian Studies (PLAAS) at the University of the Western Cape is among those who strongly advocate a shift away from support for large-scale commercial farming to smallholders. According to Cousins,[46] 'mentors' were 'milking' the projects. Government is considering land ceilings (in terms of the Regulation of Land Holdings Bill) in spite of evidence of their ineffectiveness in countries where they have been introduced [47]. Questions therefore continue to abound regarding the effectiveness of financial support to small farmers, and how to provide effective post-settlement support.

Meanwhile livelihoods-related research has revealed that improved access to basic services (water, electricity and sanitation) in poor areas has had a significant impact on quality of life, while the greatest impact on poverty so far has been the introduction of social grants, of which over 16 million South Africans are recipients [48]. Other protection services such as provision of food, increasing subsistence production and strategies to moderate food prices have also reportedly made a significant impact [49].

3. Rural market-based investments

Market-based investments in the communal areas are constrained by weak or non-existent land administration institutions [50]. Direct appropriation by private investors is not possible, given that formal ownership is in the name of the state. There are nevertheless disturbing patterns of indirect appropriation by mining companies in the mineral-rich communal areas. Traditional authority structures provide a vehicle for private mining corporations to strike deals that do not benefit the rights holders, whilst creating wealth for corporatised structures directly connected with the chiefly networks [51]. There has been increasing resistance from local communities to this form of land rights dispossession, resulting in several court cases. In some communities there has been violence and even assassinations [52].

The government has drawn up a controversial Mining Charter, which appears to be more concerned about changing the racial structure of the ownership of the mines (which are still largely white-owned) through mandatory shareholding by blacks, rather with upholding the rights of community landholders to share in the benefits [53].

Some novel land restitution settlements have occurred on land with eco-tourism/game lodge enterprises, whereby successful restitution claimants who have become the owners of the land have agreed to remain off the land in return for shareholding in the businesses. These projects are, however, haunted by contestations between traditional authority and community structures over control of the land and income, as well as the power dynamics between owners and entrepreneurs [54].

4. Urban Tenure

Urban land tenure and the release of land for low-income families is entangled with housing policies. The state has invested a great deal in subsidised housing in urban areas for low-income families. Section 26 of the Constitution states in 26(1) that “[e]veryone has the right to have access to adequate housing” and 26 (2) "[t]he state must take reasonable legislative and other measures, within its available resources, to achieve the progressive realisation of this right". South Africans born both before, as well as after, the end of apartheid have interpreted this to mean “they have a right to a free house.” Indeed the policy approach has been to attempt to fill the gap by having all expectant beneficiaries listed on housing waiting lists. The result has been the inevitable preoccupation with clearing the housing 'backlogs' and eliminating housing shack settlements by providing mostly free-standing and usually low quality housing.

In spite of an estimated R125 billion spent [55], housing lists and informal settlements are growing apace —the latter estimated to have grown by 800% since 1994 [56]. The state housing schemes simply cannot keep up with the rate of urbanisation along with new and smaller family formations, population growth and inwards migration. The approach has created a seemingly bottomless pit of housing 'backlogs.'

Since 1994, more than 3 million housing units (including both subsidised and rental housing) have been delivered to poor and low-income households. South Africa nevertheless sits today with a housing backlog of 2.1 million units. How can this be? More precisely, how can South Africa have spent R125 billion, in 2010 prices, over 20 years, delivered more than 3 million units, and yet have a larger housing backlog than when it began?[57]

The Minister of Human Settlements has recently proposed a more realistic approach in recommending releasing land for low-income urban residential development with provision of basic planning, infrastructure and services, while allowing residents to build their own houses with state assistance for the indigent [58].

Housing policies have undergone numerous twists and turns over the past twenty years, from favouring only formalisation to later recognising the need for in situ upgrading of informal settlements [59]. For the latter to be effective and to gain traction, however, the state would need to be committed to more complex and qualitative engagement with local realities.

The Minister of Human Settlements has recently acknowledged the conundrum facing housing delivery. The Minister stated "South Africa is an undisputed world first in its delivery of housing for the poor. However, we as a people have to rediscover the distinction between hope and expectation" [60].

With regard to the tenure of houses, the government continues to rely on polices that encourage formality and title deeds, rather than looking for alternatives. There is evidence that at least 50% of the titles to the new houses are not registered in the Deeds Office which points to systemic problems with the titling policies [61]. There is a vibrant market in 'informal' buying and selling of both shacks in the so-called informal housing sector, and also subsidised houses in formalised contexts. The prevailing practices for transferring shacks and houses involve locally drawn up systems of recordal, e.g. widespread use of affidavits combined with local forms of witnessing (e.g. by the police and/or local civic organisation) rather than registration in the Deeds office [62].

Urban Land rights advocates propose approaches that are being tested in various pilot cases [63]. These idea such as settlement-wide zoning using existing planning laws and in situ incremental upgrading using local land records systems to record rights.

While land rights legislation has ameliorated eviction and dispossession, it has not resulted in strong legal recognition of off-register rights. Most of the black population, representing about two thirds of the country's population of 56 million, still have off-register land right that, though protected by the laws mentioned in section 2 above, are not fully legally recognised by any system of national recordal, and lack detailed legal content. There is no land administration system to manage and administer them, other than the proposed and controversial Communal Land Tenure Bill and the equally controversial Ingonyama Trust that owns and administers customary land in Kwazulu-Natal [69].  These do not in any case include the millions of people in informal settlements and on farms.

Though there are strong locally legitimate customary practices in rural areas, sometimes referred to as 'living law', from the point of view of state governance, these rights lack full legal status in terms of property law. They tend to be minimised or glossed (and thus masked) by terms such as 'informal' and 'communal' tenure. Powerful commercial and traditional leadership interests easily trump them. The latter continue to exert a powerful hold over patronage networks in rural areas, and in some cases act as unofficial agents to mining corporations.

There has been a woefully inadequate response to recognising rights through recordal systems in expanding informal settlements in urban areas, and these rights remain insecure and subject to distant housing or settlement formalisation proposals. On farms, the rights of black occupier and worker families continue to be highly insecure and contingent. Farmdwellers are increasingly understanding and pressing their legal rights as contained in ESTA, and there have been various favourable court judgements in eviction cases, but these victories do not necessarily have a systematic impact on farmdwellers' rights. There is a lot of anecdotal evidence that ESTA has had an unfortunate backlash in motivating farm owners to evict or displace farmdwellers, but this cannot be verified through research findings at this stage.

State and private sector policy pundits continue to push for titling. Critics propose an alternative understanding of, and approach to, legal recognition of tenure in South Africa. A new publication on tenure security outlines the flawed approach to tenure reform via full conversion to title, presenting evidence from a wide range of case studies on tenure practices on the ground. The authors conclude that assumptions underlying current tenure policy, such as for example, the proposals in the form of CLTB and urban titling, are both unrealistic and unjust, and propose an alternative conceptualisation of off-register tenures as 'social tenures' that can be recognised in their own right without full conversion [70].

In order to strengthen existing land rights, advocates propose the strengthening of the current protective land rights legislation and the development of land administration systems to back them. For example, the Interim Protection of Informal Land Rights Act (IPILRA) should be made permanent, while others amended in various ways to provide for stronger rights. However, these rights are to be supported by a strong land administration system [71].

Some land administration policy experts propose that there should be a countrywide system of land administration with infrastructure to support a 'land records' system that is suitable for both urban and rural contexts where the recorded evidence is needed for legal recognition that differs from the Deeds registry system [72]. This will shift the focus to rebuilding land administration institutions to enumerate, adjudicate, record and maintain recorded rights, as well as resolve disputes. Although the records will be held locally, all aspects of the system should be regulated by national norms and standards, as well as Constitutional principles of equity, including gender equity, with the ultimate goal of integration. This approach will shift the emphasis from a singular deeds registration system, towards a system that can accommodate pluralism and thus also work for the majority of land rights holders who live on farms, in informal settlements and in communal areas. The Deed Registry would be retained as the repository for title deeds, but expanded to function as the overall custodian of the land administration system of the country as a whole. This will require incremental adjustments with the long-term goal of institutional integration and accommodation of diversity [73].

These proposals form part and parcel of much wider initiative under the Collaboration on Alternatives for Land Policy and Law that has produced a large volume of material to the High Level Panel of the South African Parliament [74], which is about to released to the public. The papers and memoranda provide evidence of the problems besetting land reform, and make recommendations for wide-ranging changes to all aspects of land reform, including restitution, redistribution, mining and tenure laws applicable to all categories of rights holders. The general approach is to call for a new White Paper on Land Reform to replace the old one of 1996, and for a complete reappraisal of all the laws from a perspective that reasserts the values and norms of the Constitution, and provide new framework legislation to ensure overall coherence.