The road to TRAs is paved with good intentions: Dispossession through delivery in post-apartheid Cape Town

Introduction

I was conducting research in a land occupation facing eviction in Mitchell’s Plain, the second largest township in Cape Town, South Africa. Five thousand ‘backyarders’ from the area – people living in shacks behind formal houses – were tired of overcrowding, or else they could no longer afford the seemingly meager rent. Despite the township’s size, it is also one of the city’s more peripherally located areas, a function of its late apartheid origins in the forced relocation of more than 60,000 so-called ‘Colored’ ¹ residents from the city centre (Western, 1997). It was in this occupation that I encountered Kyla, who had been struggling against the Anti-Land Invasion Unit and the police, not to mention constantly travelling to the city centre for court hearings, to defend her place on this open field for four months by the time I met her. She recalled how one morning, an employee from Cape Town’s Department of Human Settlements came and talked to a group of squatters: ‘They told us about Blikkiesdorp, that we’d have nice houses and schools’.

Blikkiesdorp, discussed in greater detail below, is one of 12 temporary relocation areas (TRAs) within Cape Town’s city limits (Pillay et al., 2017), all created since the early 2000s as presumably temporary housing. The liberalisation of apartheid in the 1980s increased the pace of black urbanisation, with the number of new informal settlements increasing more than seven-fold since the transition (Tomlinson, 2015), leaving more than a fifth of Capetonians living in shacks according to the latest census. When residents face crises such as shack fires or flooding, or else when they illegally occupy land to erect new homes, they are frequently offered a spot in a TRA. In compliance with a legal mandate that municipalities must offer ‘alternative accommodation’ in cases of eviction (Huchzermeyer, 2003a), Cape Town’s Department of Human Settlements has increasingly used TRAs as ‘alternative accommodation’ in recurrent mass eviction cases rather than as temporary shelter in instances of unpredictable environmental disasters. I learned of no case in which residents moved to a TRA from a land occupation were subsequently moved elsewhere. Indeed, as one of the City’s chief urban designers told me, ‘These TRAs that we’re seeing everywhere – they’re not temporary’. Instead, residents remain trapped in a state of permanent temporariness (Oldfield and Greyling, 2015; Yiftachel, 2009a, 2009b), having received ‘alternative accommodation’ but finding themselves torn from the networks that were at the basis of their strategies of reproduction. These strategies and populations are ‘concurrently tolerated and condemned’, as Yiftachel (2009a: 90) puts it, ‘perpetually waiting “to be corrected”’.

TRAs are poorly located, with the majority situated in the apartheid-era blackbelt above Mitchell’s Plain. Residents lose community support networks and often find themselves in unfamiliar gang terrain (Pillay et al., 2017). TRAs reinforce the spatial mismatch wrought by apartheid, with racialised populations moved further from the city centre, increasing their transport costs while decreasing their likelihood of employment (Huchzermeyer, 2003b; Macgregor et al., 2007; Turok, 2001). They are often far from social and municipal services. This was Kyla’s experience, as she continued to explain to me why she would never accept alternative accommodation in Blikkiesdorp: ‘There’s not even a day clinic there, and the hospital is too far. The school is opposite the freeway, and children die crossing. We must bring the Human Rights Commission into this’.

From a policy perspective, the disjuncture between the City’s perception of TRAs as sufficient ‘alternative accommodation’ and Kyla’s understanding of Blikkiesdorp as dispossession is an instance of what Watson (2003) calls ‘conflicting rationalities’. She argues that ‘a vast gap exists between the notion of “proper” communities held by most planners and administrators’ and the quotidian survival strategies of residents themselves (Watson, 2003: 401). Along similar lines, Charlton (2009) labels this incongruence ‘competing rationalities’, pointing out that the understanding of ‘success’ varies depending upon viewpoint. While the national government may understand housing distribution in terms of ‘poverty alleviation’, ‘urban citizenship’, or ‘enabl[ing] housing markets to work’ (Charlton, 2009: 303), at the level of the household it may be experienced as social dislocation.

More broadly, I argue that this disjuncture represents a novel mode of state-induced socio-spatial dislocation, or what I call dispossession through delivery. I understand ‘dispossession’ as the physical separation of residents from their homes, land and social networks. Relocation constitutes a moment of dispossession insofar as residents’ new homes are divorced from established networks, lack access to expected services and are further from employment opportunities. In the next section, I demonstrate how in contemporary writings on dispossession, the phenomenon is typically understood as an effect or component of neoliberalism, with social spending understood as its antidote. I complicate this antithesis by drawing upon recent work on South African cities that questions the utility of these ideal types, but that deliberately refrains from suggesting an alternative characterisation. Instead, I take the ‘conflicting rationalities’ literature to its logical conclusion, positing dispossession through delivery as an alternative to either/or understandings of neoliberalism and social spending. I then discuss the origins of the TRA model and how it came to function as a technology of dispossession, before concluding with an extension of the concept to other housing programmes in Cape Town, including informal settlement upgrading and the distribution of state-provisioned formal housing. Rather than multiple simultaneous ‘cases’ of dispossession through delivery, I explain how this concept requires us to think TRAs, RDP houses, site-and-services ‘housing opportunities’, backyarding, informal settlements and land occupations as comprising a singular articulation with multiple sites of dis- and re-possession.

Beyond either/or

In the first decade of the 2000s, a substantial assortment of social scientists viewed South African social policy in the 1990s as a textbook transition to neoliberalism. After a brief period of social spending defined by the African National Congress’s Reconstruction and Development Program (RDP), strategically designed to counteract inequalities wrought by centuries of colonialism, segregation and above all apartheid, South Africa underwent its neoliberal turn (Alexander, 2002; Bond, 2000a; Marais, 2011; Satgar, 2008; Saul and Bond, 2014; Seekings and Nattrass, 2006; Terreblanche, 2002). Within two years the RDP office was closed, and the ANC capitulated to the imperatives of transnational capital, implementing a new macro-economic strategy ironically named Growth, Employment and Redistribution (GEAR) in 1996. GEAR was represented as a programme of indiscriminate neoliberalisation, characterised by successive rounds of privatisation (Bond, 2000b; McDonald and Pape, 2002; Miraftab, 2004) and the requisite wave of anti-privatisation movements launched in response (Buhlungu, 2006; Runciman, 2012; Veriava and Naidoo, 2008). Analysts emphasised the role of the World Bank and International Monetary Fund in cutting services and implementing user fees (Bond, 2003; cf. Gilbert, 2002) and the Finance Ministry’s liberalisation of capital controls (Ashman and Fine, 2013; Ashman et al., 2011).

Indeed, there is something in this set of texts that recalls what Hart (2018: 373) recently described as ‘a form of comparison that asserts a general or encompassing process, and then considers specific “cases” as variants of that process’. Typically ‘coded as “Marxist’” (Hart, 2018), this approach to understanding capitalist development refuses to consider conjunctural trajectories, but instead identifies an ever growing number of ‘cases’ captured by a monolithic global capital, remaking the entire globe in its image.

This sweeping approach to neoliberalism has more recently given way to its antithesis: writers from disparate political positions and disciplines have contested the claim that the South African pendulum swung from social spending to neoliberalism, instead insisting that upon empirical examination, the welfare apparatus actually expanded over the period in question (Breckenridge, 2014; Ferguson, 2015; Habib, 2013; Harris and Scully, 2015; Seekings and Nattrass, 2015). This reflects a broader tendency of augmented social spending across middle-income Southern countries over the past two decades (Burdick et al., 2008; Harris and Scully, 2015; Sandbrook et al., 2007). Whether in terms of cash disbursement or the distribution of free formal housing, social spending was not haphazardly slashed over the decade following the passage of GEAR, but actually increased.

Some scholars of the Southern welfare state deploy the concept of decommodification to capture the effects of social spending at work. Seekings and Nattrass (2015: 133) argue that ‘stark decommodification’ underpinned the project of the welfare state, in terms of both contributory insurance schemes and social assistance programmes. Increased social expenditure constitutes an ‘unprecedented decommodification of the poor’ (Seekings and Nattrass, 2015: 135), meaning that residents no longer have to depend upon wages to reproduce their own livelihoods. While intuitively it might make sense to conceive of decommodification as a general tendency, it loses its specificity when the authors describe housing delivery as ‘a mix of “decommodification” and “commodification”’ (Seekings and Nattrass, 2015: 183). Given that state-provisioned housing is constructed by private contractors and is formally titled, in no sense is housing actually decommodified other than the fact that the state subsidises its construction. Like Seekings and Nattrass, Harris and Scully (2015) describe decommodification as a state-implemented process, and they define decommodification as the reduction of dependence upon wage labour, asserting that, ‘decommodification of livelihoods and protection from market-based dislocations are becoming firmly entrenched norms among their citizens’ (Harris and Scully, 2015: 440).

In all of these accounts, neoliberalism and social spending remain necessarily antithetical. Social expenditure, including cash disbursements, state subsidies and service and housing delivery, counters the dispossession wrought by the neoliberal market. If Harvey (2003) once described the neoliberal regime of accumulation as rooted in dispossession, we might characterise its corrective – social spending – as a means of repossession: decommodification entails the non-market mediated provisioning of goods and services necessary for sustaining livelihoods, reversing the destruction wrought by dispossession. But when dispossession and social spending proceed in tandem, as I argue they do in post-apartheid cities, neither narrative can adequately capture the nature of this process.

When we only have two poles between which we can position ourselves, any instance that does not conform to either ideal type must be described as a hybrid of both. Lemanski describes Cape Town as a combination of ‘pro-growth and pro-poor strategies’ (Lemanski, 2007: 454; cf. Houghton, 2013). She points out that while national government initiated the devolution of economic growth strategies onto municipalities (cf. Rogerson, 2003), Cape Town built redistributive programmes into these measures (cf. Parnell et al., 2005; Rogerson, 2000), subsequently describing this as ‘neoliberalization alongside a pro-poor agenda’ (Lemanski, 2011: 58). Seekings and Nattrass describe this hybrid approach using analytic constructs such as ‘semi-social insurance’ and ‘semi-Keynesian macreconomic policies’ (Seekings and Nattrass, 2015: 141, 19), proceeding to describe jobs as ‘work opportunities’, workfare as a redistributive programme and sweat equity as housing provision (Seekings and Nattrass, 2015: 143–145, 183). Again, they describe policy programmes in South Africa as a combination of the two ideal types.

Ferguson (2010) suggests that our thinking remains mired in mid-20th century European models. ‘Couldn’t the Democratic Republic of the Congo solve all its problems, reformers seems [sic] to say, if only its government would start to behave like that of Sweden?’, he asks facetiously. ‘Well, maybe so, but it’s the sort of ahistorical and asociological formulation that is worse than useless’ (Ferguson, 2010: 168). Indeed, the rise in social spending in Southern cities amid a backdrop of neoliberal macroeconomic policies should not be taken as a paradox, but instead must be read as an outcome of ‘the contingencies of democratic politics’ (Ferguson, 2010: 171) – and, I might add, as outcomes of postcolonial and post-authoritarian transitions.

This is where considering conjunctural political context proves essential. As Parnell and Robinson emphasise in their critique of even the most nuanced theories of neoliberalisation, in the case of South Africa, ‘the relevant time frame is not a post-1970s crisis, but a post-1994 experience of state restructuring’ (Parnell and Robinson, 2012: 599). When we conceive of housing provision, for example, as a post-apartheid (or even postcolonial) remedial project, the context of democratisiation, not to mention the assumption of political power by liberation movements (Southall, 2013), plays a key role in tying social spending to the legitimacy of new regimes. On the other hand, I wouldn’t go as far as these authors do and substitute a ‘practice-generated literature’ for an account of the development of social spending in a neoliberal context. In the case of housing policy, it was a classically neoliberal think tank, the Urban Foundation, that played the dominant role in transitional housing discussions (Bond, 2000a). And turning to policy literature rather than pairing it with ethnographies of policy-as-implemented fosters an approach that takes pro-poor policy at its word. Instead, these two approaches always need to be held in relation to one another. But Parnell and Robinson’s point is well taken: reading Southern cities as proliferating variations of ideal types generated from Northern cases inhibits our ability to observe novel phenomena, and therefore to theorise from Southern cases (Parnell and Robinson, 2012: 596; cf. Roy, 2009).

The introduction to one influential volume on neoliberal urban politics makes precisely this point: if ‘neoliberalism … has been transformed to the point where it is no longer recognizable as such’ (Leitner et al., 2007: 10), the concept may inhibit, rather than enable, the understanding of urban policy in particular contexts. Instead, as I argue here, we must move beyond simply trying to locate each city on a spectrum ranging from neoliberal to social democratic, but instead devise novel concepts for understanding how urban policies play out in practice. While dispossession through delivery is hardly sufficient to capture the entirety of post-apartheid urban change, it is a first stab at making sense of unanticipated pairings. In much the same way, Parnell and Robinson (2012: 603) ask us to consider the unexpectedness of a social safety net coming from above, or civil society demands for state abdication. In what follows, I develop the concept of dispossession through delivery through the emergence of Cape Town’s TRA programme. But first, I situate this ‘emergency housing’ in relation to the broader processes of urban change in which it emerged.

Managing an emergency

Despite the distribution of millions of subsidised formal homes since 1994, the number of residents living in shacks has actually increased, both nationally and in the Western Cape, and no major municipality has succeeded in substantially reducing its housing backlog. Apartheid meant the repressive fixing of unwanted populations in peripheral areas, typically lacking adequate employment, infrastructure or education systems. These areas were known as ‘Bantustans’ and were created by the apartheid government as ethnically homogenous, autonomous territories for ‘African’ populations. Far from designating the autochthonous population, ‘African’ was deployed by the apartheid state as a general descriptor for all speakers of Bantu languages; the so-called ‘Colored’ population, who predominantly speak a dialect of Afrikaans, were excluded from this category. The government strategically disaggregated ‘African’ into its constituent ethnicities, with each assigned its own Bantustan. In the case of Cape Town, isiXhosa-speaking residents were deported to either the Ciskei or the Transkei in the present day Eastern Cape. As mobility restrictions were lifted in 1986, black rural populations flowed into cities en masse – above all, in search of decent employment opportunities.

In addition to in-migration from the Eastern Cape, the proliferation of informal housing after 1986 was even more so the consequence of migration within the city. As Makhulu (2015: 175n.4) observes:

much of the history of movement and mobility within the Cape Flats is a history of internal displacements – a history very often occluded by the significance afforded forced removals of colored communities from the City Bowl as well as the histories of migrants from the rural Eastern Cape.

Department of Human Settlements officials I interviewed repeatedly deployed narratives equating all squatters with returnees from former Bantustans, ignoring intra-urban migration altogether. The City’s head of land acquisitions told me that land occupiers ‘perceive Cape Town’s level of better services for most people, and particularly for people from the Eastern Cape … Cape Town is a desired destination’. She proceeded to refer to participants in land occupations as ‘migrants’.

By contrast, most of the participants in land occupations that I came to know over the course of my fieldwork grew up in the same townships in which they later built shacks. Typically, their parents or grandparents received formal housing in newly constructed townships after being forcibly removed from the city centre in accordance with the apartheid state’s racialised zoning of the city (Western, 1997). Despite having access to electricity, running water and private bathrooms, the severity of overcrowding impinged upon residents’ sense of autonomy and dignity (Lemanski and Oldfield, 2009: 638). In one home in which I stayed during fieldwork, Fatima and Mustafa, both in their mid-60s, lived with two of their five adult children in a modest three-bedroom home. Their daughter Aisha lived there as well with her third husband, Muhammad. She had two children with her first husband and two with her second, and all four lived in the house. Muhammad’s two teenage sons would also sometimes stay there. In addition, they effectively adopted their homeless friend’s four-year-old daughter, and her mother would stay with them for extended periods. The consequence was nearly a dozen people in a 40 m² home, meaning no privacy whatsoever, and certainly no sense of an immediate family separate from an extended one. While it may seem counterintuitive in terms of the quality of one’s home or in terms of security to participate in a land occupation, newly constructed shacks are, as Makhulu (2015: 53–54) aptly describes, ‘associated with a sense of emancipation, both affective and political. [They are], for young lovers at least, a space apart from the domestic scene controlled by parents in which gender and generational customs might be suspended’. When I asked Aisha why she would risk participating in a land occupation if she had a bed, however crowded, in her parents’ house, she insisted, ‘I wanted a home of my own, a home for my children, for my family’. For her family then, this was about securing autonomy and dignity.

Land occupations as a form of internal migration, in combination with in-migration from former Bantustans, meant that demand for new housing vastly outstripped supply, and that the post-apartheid housing distribution programme was doomed from the get-go. The South African government distributed over 2.5 million formal homes and 1.2 million serviced sites since 1994, and plans for an additional 1.5 million by 2019 (Tomlinson, 2015). Using the census’ average households size of 3.1, this means that more than a fifth of South Africans were housed under this programme. Yet, during the same period the number of new land occupations has increased, and the number of informal settlements has grown by 650 percent (Tomlinson, 2015).

In any postcolonial situation in which restrictions on mobility are suddenly lifted, rapid urbanisation means that remedial social spending can only do so much to alleviate persistent socio-spatial inequalities. Given this realisation, by the early 2000s, South African municipalities began to implement a series of stopgap measures. Foremost among these were TRAs, designed to provide transitional shelter while residents waited for more durable housing. Emerging over the course of the early 2000s in most major South African municipalities, they quickly degenerated into stigmatised zones of exclusion, as there proved to be nothing temporary about them. In Cape Town, ‘temporary’ units are still standing more than a dozen years after their initial construction. The most notorious of these is Symphony Way TRA, colloquially known as Blikkiesdorp, Afrikaans for ‘tin can town’, after its rows of thin-walled zinc structures. Blikkiesdorp has come to function as a metonym of sorts for the state’s larger TRA programme.

TRAs potentially exacerbate social isolation and atomisation (Macgregor et al., 2007; Pillay et al., 2017), as well as failing to actually serve as transitional means towards more adequate housing (Chance, 2015; Ranslem, 2015). Rather, they reinscribe squatting as an adequate form of shelter, accompanying a larger shift in South African housing policy from actual housing delivery to self-help housing schemes. In the case of TRAs then, social spending and dispossession are not necessarily antithetical. Indeed, it is precisely the delivery of precarious housing in far-flung sites that constitutes the moment of dispossession: dispossession through delivery. While conditions in land occupations should not be romanticised, as they frequently remain sites of extreme destitution and violence, they simultaneously represent emancipation from the overcrowded housing of participants’ parents and grandparents and from ‘appalling backyard conditions and exploitative landlords’ (Lemanski, 2009: 474). Land occupations are a means of establishing one’s own household, and therefore represent all of the dignity and autonomy bound up with the reproduction of the immediate family (Lemanski and Oldfield, 2009; Makhulu, 2015; Tomlinson, 2006). Under the sign of the provision of housing and social services, TRAs represent the denial of this emancipation and the separation from the material means of achieving it. Dispossession need not be the product of malicious intentions or a deliberate programme of accumulation; indeed, as TRAs demonstrate, it may be a central component of social provisioning itself. The concept of ‘dispossession through delivery’ forces us to think through alternative strategies of decommodification that circumvent social dislocation.

A brief history of TRAs

Rather than a conscious strategy of remedying the lack of capacity of the delivery apparatus, the building of TRAs was a haphazard development over the course of the early 2000s in numerous South African cities. In Durban, TRAs were called ‘transit camps’ and were tied to a range of rationales, providing alternative housing for victims of shack fires or those displaced by new highway construction, as well as for those living in unregistered informal settlements (Hunter, 2012; Hunter and Posel, 2012). These unregistered settlements – called ‘land invasions’ by the state – were viewed as both a human rights emergency as well as a blight on the city’s status as ‘world-class’. In Johannesburg, TRAs went under the name ‘decant camps’, and they were frequently deployed in situations of environmental hazard (Ramutsindela, 2002; Ranslem, 2015). Most commonly, new land occupations took place on land prone to dolomitic sinkholes, and safeguarding against potential collapse would require the installation of expensive concrete slabs under all new shack settlements.

In Cape Town, the story is roughly similar. The first TRAs were constructed in the early 2000s for residents of land occupations deemed too overcrowded or located on unsuitable land. Invariably residents were relocated to peri-urban sites far from the central business district, increasing not only commute time but also taxi fares to the city centre for work. Blikkiesdorp was built following a massive shack fire in Langa, one of the former black townships closest to the city centre. In 2005, this fire destroyed the homes of over 10,000 residents of an informal settlement in Langa called Joe Slovo. Many of them were initially housed in a TRA in the same township called Langa Intersite, but provisions were only made for a fraction of the displaced residents. The idea floated by the Department of Human Settlements at the time was that the provincial government would construct new apartment blocks for those who lost their homes, and they would recover costs by renting out additional new apartments to those who could pay. After years of delays, new housing was only produced for a small percentage of displaced residents, with the large majority remaining without options. Given that the TRA at Langa Intersite was already full, the Department of Human Settlements built a new TRA 20 kilometres further from the city centre on land temporarily leased from the airport. Today it is the most notorious TRA in the country and houses roughly 1800 structures.

Nearly a decade after Blikkiesdorp’s initial construction, there are no plans for the relocation of its population. The TRA was intended to last for roughly six months as a stopgap measure in the emergency situation that was the Joe Slovo shack fire. But few if any TRA structures have served this transitional function. Instead, TRAs have become regularised as a form of state-provisioned housing even in non-emergency situations. Despite their piecemeal emergence, TRAs have been coordinated after the fact under the Emergency Housing Programme (Levenson, 2017a).

The enshrinement of TRAs in official housing policy signals a major shift in their intended function. What began as temporary solutions to environmental crises and natural disasters soon became routine, state-provisioned housing options in cases of effective homelessness. When residents with few other options would band together and coordinate mass land occupations, building informal housing on empty plots of land, the City typically viewed this as a threat. Indeed, around the same time Blikkiesdorp was constructed, the City of Cape Town convened an official Anti-Land Invasion Unit as part of its Department of Human Settlements. This Unit worked in close coordination with police to surveil and ultimately eradicate newly emergent informal settlements.

So-called ‘land invasions’ were construed as threats by Department of Human Settlements officials, who argued that they undermined the muncipality’s system of housing distribution. Residents in need were expected to wait patiently for state-provisioned housing, even when this took decades, as it normally did, and even when there was no guarantee that residents would even receive formal housing. This uncertainty recalls sociologist Auyero’s (2012) characterisation of Argentinian welfare recipients as ‘patients of the state’: they are stripped of their agency, delegating all of it to the state as they wait endlessly, becoming objects of governmental calculations. Yet as Oldfield and Greyling (2015: 1100) demonstrate in their study of the South African housing waiting list, waiting also ‘shapes a politics of quiet encroachment “in the meanwhile”’, and more specifically, as I argue here, generates land occupations.

In Cape Town, housing officials routinely referred to participants in these occupations as ‘queue jumpers’, implying that the entire point of living in a shack settlement is to bypass the waiting list for state housing and jump to the front of the queue. Their proper place is as data in a formally rational system of housing distribution that includes a waiting list – the ‘queue’ in question. Marlize Odendaal with the City’s Department of Human Settlements was in charge of land acquisition, the purchasing of urban land for resettlement and housing distribution. When I interviewed her in 2013, she told me:

From my perspective I think a lot of what is happening in terms of land invasions is need-driven … But having said that, urbanization alone is a reality that we need to cope with, but I think a lot of it is politically motivated and purely aimed at embarrassing and/or just jumping queue.

She frames participants as antagonists of a benevolent welfare apparatus, as opportunistic actors who undermine the formally rational logic of the housing waiting list. This logic of perceived adversariality characterised many of the interviews I conducted with DHS officials. The head of the City’s Anti-Land Invasion Unit, Stephen Hayward, provided me with his take on occupations:

The City was losing a lot of land because of people illegally invading. And the fact is that the City was losing the land and couldn’t plan and continue the process of housing. You would have people that would settle in a low-lying area where every winter it gets flooded. And every winter the people complained and demonstrated because the City wasn’t looking after them. So the City would build housing quite close by and have to move them there.

Hayward here opposes the government’s ability to distribute housing on the one hand, and new land occupations on the other. This governmental rationality is premised upon a scalable model of housing provision. It requires ordered populations able to wait until housing supply can meet a ballooning residential demand. When they cannot, they are framed as threats. But these objects of governmental calculation have a very different operating logic: if they are perpetually evicted from land occupations as they wait for government housing, where are they supposed to go in the meantime?

Meanwhile, a series of South African Constitutional Court rulings provided further justification to the shift in TRAs’ function from emergency measures to last-resort housing provision. Years before the construction of Blikkiesdorp, the Constitutional Court issued one of its most famous decisions in the 2000 Grootboom case. This ruling required municipalities to provide ‘alternative accommodation’ in cases involving ‘the immediate needs of those living in intolerable conditions’ (Huchzermeyer, 2003a: 81). This requirement was reiterated in a 2011 ruling that upheld municipalities’ responsibility to provide alternative housing in cases of eviction, even in cases involving occupied private property (Strauss and Liebenberg, 2014; Wilson, 2011). TRAs were included in the category of legally legitimate ‘alternative accommodation’. When multilateral institutions like the World Bank and UN-Habitat praised South Africa’s ‘alternative accommodation’ provision as in line with its Constitutional guarantee of the ‘progressive realisation’ of the ‘right to have access to adequate housing’ (RSA, 1996), they were including TRAs in this definition of ‘adequate housing’. But from the perspective of residents, there was nothing adequate about it.

Why would anyone refuse adequate housing?

If TRAs are defined as adequate ‘alternative accommodation’ and praised as a viable stopgap measure, then why do these not constitute a substantial counterweight to proliferating land occupations? From the perspective of many housing officials, this is a perfectly reasonable view. As Blikkiesdorp came to function not just as emergency housing, but as a potential option for land occupiers facing eviction, it was commonly described to me by Department of Human Settlements employees as a community of homes built and distributed free of charge – almost as if receipt of a zinc shack were the equivalent of receiving a formal, multi-room house. Officials frequently expressed bewilderment that residents would not simply accept these handouts, failing to comprehend the deeply entrenched aversion to the camp. Fear aside, relocation to Blikkiesdorp typically meant leaving one’s longstanding community, including employment opportunities, and adding substantial time (and therefore transport cost) to the commute to any work in the central business district.

But fear was also a huge aside. When I first started visiting residents in Blikkiesdorp in 2012, my contacts in land occupations and other informal settlements would insist that I should not be going, even when driving in broad daylight. The general perception of Blikkiesdorp among Cape Flats residents, even those who lived in statistically more dangerous neighbourhoods, was that it was an unbridled warzone. Indeed, it was not untrue that the place was marked by extreme violence. The recent turn towards ad hoc relocations to Blikkiesdorp resulted in gang members from rival territories being placed in close proximity. This partially explains other residents’ fear of the TRA: even where violent crime levels are higher, the perpetrators are their perpetrators, affiliated to a field of gangs that they can actually comprehend. By contrast, Blikkiesdorp was perceived as particularly chaotic because no major gang was able to control a sizeable swathe of territory; instead, fractions of rival territories abutted each other without any clear lines of demarcation.

On numerous site visits, residents told me tales of break-ins in which gang members tore through their thin zinc walls with sharp blades. One resident named Auntie Jo told me that often these were not even break-ins, but simply gangsters fleeing the police. Rather than be caught out in the wide dirt alleys between blikkies, it was safer for those fleeing the police to move through houses than between them.

On another occasion, a resident named Mary had her home demolished by a group of gang members for beginning to organise against murders on her block. The first time I visited her, we arrived to find her blikkie a mess. Mary’s 15-year-old son approached us and explained that police had visited the house earlier that day, beating him and the other children inside with the flat side of a shovel. The police kept repeating, ‘Where’s the dagga [marijuana]?’, he told us. The two officers kept insisting that the children were hiding drugs. Mary then sent her six-year-old son to the corner store to buy us a bottle of soda. He returned minutes later in tears. Two gangsters had robbed him of his 10-rand note at gunpoint. When she used to live in a nearby working-class area called Belhar, Mary told me, she would have marched right outside with her child and confronted the gang leader for allowing one of his footsoldiers to harass a neighbour. But now, she reiterated, she had no idea who controlled the stick-up operation. She no longer knew who to confront in such a situation.

Occasionally officials would acknowledge that fear was a substantial factor, even if they claimed not to agree with this sentiment. Hayward of the Anti-Land Invasion Unit told me about one group of squatters that he tried to convince to move to Blikkiesdorp the previous year. They had organised a land occupation on a field called Kapteinsklip but were subsequently evicted. In accordance with ‘alternative accommodation’ requirements, they were offered the Blikkiesdorp option before being left to fend for themselves. He told me:

In Kapteinsklip, we didn’t have to because the court said that we didn’t have to provide alternative accommodation. But we dealt with the Legal Resources Centre, and they said, ‘Have pity on the people’. So we negotiated, and we eventually found space for them at Blikkiesdorp … Two families accepted, [and] we actually went to pick them up, [and] took them to Blikkiesdorp to show them what it looks like … When they got there, the people said, ‘We’re waiting for you to come. We’re going to rape and kill you’… So they are all too scared to go there. That’s what they tell us.

In Hayward’s characterisation, this was not even about complying with the legal mandate handed down by the Constitutional Court, but rather about the benevolence of the post-apartheid welfare state: ‘hav[ing] pity on the people’. But in either case, what was represented as beneficent social provision proved to be the least desirable option, with destructive consequences.

I previously conducted an interview with a married couple who were part of the occupation. They were among the group that the ALIU transported to Blikkiesdorp in order to convince them to accept housing there. The husband, an ex-gang member himself and a lifelong resident of Mitchell’s Plain, told me:

They explained to us how okay Blikkiesdorp is, and when we got there, the gangsters that was standing there was saying, ‘If you gonna come live here, we gonna kill you. We gonna rape your women’. And then the cops didn’t want to get out of the van. But they want us to live there! And then we spoke to the guy, the main guy there that works for the City, that sees to the maintenance and all that. We asked him where’s the school. And he kept on saying, the school is here, there’s the school. And we keep asking, where’s the school, and he keeps on showing us, the school is right here. I can’t see the school. And he said, no, it’s down that road and down that road, and I told him, that’s not what you said in Mitchell’s Plain. You said the school is right there, and it’s not right there! That’s why we can’t move here. And at the same time, some lady comes there and says, they just cut through her hokkie [cage or pen, as for animals] with a scissor and stole her DVD. And then we told that guy, while we here now for 10 minutes, all of this is happening. How do you want –

His wife interjected, ‘And there was a fight!’. He continued:

People were stabbing each other, and we said, how can we come live here? Is this what is happening here? … You’re gonna see kak [shit] when you come here, and all these kind of threats. And we said we can’t, we don’t want – and since that day we decided no Blikkiesdorp.

These conflicting accounts illustrate the substantial disjuncture between local housing officials’ perceptions of their own role in managing the state’s welfare functions, and residents’ understanding of relocation as a form of dispossession. When housing officials reassert the benevolent nature of local government, they have the bigger picture in view. They recall the larger system of housing distribution backed by a legal system nearly unique in guaranteeing socio-economic rights. They also know the practical difficulties that planning entails when the parts that must be coordinated are constantly in motion.

By contrast, residents – the objects of governmental calculation – think not in terms of scalable planning models or quantifiable measures of success, but in terms of how these policies impact their everyday lives. What Hayward describes as finding space for people in need was experienced by these same residents as coerced relegation to a far-flung site lacking basic social services. It required them to leave the social networks that provided the familiarity they required to navigate life in a dangerous neighbourhood and to move to a place in which they feared that they would be stigmatised as outsiders. This logic of experienced dispossession is why the same TRA programme that could qualify as ‘alternative accommodation’ could be viewed by some residents as a threat to their very livelihoods. Now we can understand why Kyla could insist that the Department of Human Settlements should be reported to the Human Rights Commission in the opening vignette: she experienced temporary relocation as violent social dislocation.

Conclusion

Blikkiesdorp – and TRAs more generally – represent an inherent limit to thinking dispossession solely in relation to neoliberalism. As I have argued here, TRAs constitute technologies of housing distribution, but they fail to function as intended, with dispossession (rather than repossession) as their standard outcome. While we might be tempted to view TRAs as a policy misstep, counterposed to the government provision of subsidised formal housing or the upgrading of informal settlements, this would be a mistake. The concept of dispossession through delivery is not a descriptive phrase limited to TRAs, but is rather the consequence of a range of technologies of delivery. Only in thinking these together, articulated as ‘housing opportunities’ in the Department of Human Settlement’s preferred terminology, can we begin to understand this concept as a potential alternative to the neoliberalism/social spending antithesis, and to disarticulate dispossession from its presumed status as the exclusive province of neoliberalisation.

A substantial body of research on formal housing distribution in South African cities has demonstrated a slightly different form of dispossession through delivery than we observe in the case of TRAs. ‘[D]elivery’, Tomlinson concludes, ‘has been viewed as entrenching the racially segregated urban form that arose under apartheid, because most of it has taken place on the urban periphery’ (Tomlinson, 2006: 101; cf. Gilbert, 2014: 258). Turok similarly argues that apartheid’s spatial legacy has been augmented, with an increasing spatial mismatch in terms of employment, as well as skyrocketing transport costs that force some families to abandon their newly obtained homes altogether (Turok, 2001: 2352; cf. Lemanski, 2011; Levenson, 2017b). Lemanski (2009) insists that a poorly orchestrated housing delivery programme has actually encouraged informality, and in the case of Cape Town, the proliferation of backyard shacks. She concludes by condemning the ‘complicity of government policies themselves in augmenting informality’ (Lemanski, 2009: 483).

If Turok describes formal housing provision as fostering a ‘dormitory status’ in former apartheid townships (Lemanski, 2009: 2372), Huchzermeyer (2001: 306, 311) criticises informal settlement upgrading policies for creating ‘dormitory developments’ and ‘toilet towns’. Despite a promised shift from greenfield to in situ upgrading (Tomlinson, 2006), nothing of the sort materialised in any major South African city, with spatial relegation the predictable consequence (Huchzermeyer, 2011).

When the receipt of housing requires relocating to sites even further from employment opportunities and leaving established social networks, it is no wonder that municipal services and housing remain the top two causes of mass protest in the country (Alexander et al., 2014; Powell et al., 2014). Indeed, rather than thinking of dispossession as something to be entirely remedied by social spending, dispossession through delivery requires us to rethink how we conceive of urban delivery programmes altogether. The mere fact of government provisioning is no longer sufficient. Indeed, as Esping-Andersen (1990) presciently observed in the instance of advanced industrial welfare states, it does not simply mitigate existing inequalities but additionally produces novel systems of stratification. Something similar appears to be at work in post-apartheid cities, and it is our challenge then to map novel socio-spatial effects, moving beyond the stale binary.