Skeletons at the feast: A review of street homelessness in South Africa and other world regions

Abstract

Homelessness on the streets has been of concern to governments and civil society for hundreds of years, and the number of homeless tends to rise when economic conditions take an adverse turn. Laying stress on questions of access to housing, livelihoods and services, this paper compares the historical causes of homelessness in Britain and Europe, India, the US and South Africa, in order to approach a better understanding of South Africa's own homelessness situation. Internationally, the key debate is whether homelessness is due to simple lack of affordable housing, or to a range of complex factors involving poverty and unemployment. The paper argues that spatial access to street livelihoods and access to the metro core zones are critical factors linking housing access to poverty economics, and it questions whether in South Africa's situation street homelessness can be eliminated in the foreseeable future.

Keywords:

• homelessness
• street people
• shack settlements
• government policy
• poverty
• housing delivery
• Britain
• Europe
• India
• United States
• South Africa

1. Introduction

Homelessness on the streets in South Africa is a slow-moving tragedy that arouses anxiety in government and civil society, and it is the more serious because it is overshadowed by the size of the population in shack housing. As unemployment has risen, larger numbers of the poor have fallen into marginality, living on social grants and dependent on temporary work; from there, many have descended into true homelessness (see Cross & Seager, this issue).

This article looks at the historical development of policy for the street homeless, with the aim of identifying ways to formulate effective policies. The following are central questions:

1. Are the main causes to be found in South Africa's housing delivery, or in a wider context of poverty determinants?

2. How can we resolve the conflict between street homeless people's need for space to live in and the cities' need for an urban environment that will promote investment?

3. How similar are the determinants of street homelessness in South Africa to those of the industrialised and industrialising countries?

4. If government anti-poverty policy relies on household accumulation strategies based on delivery of secure housing, can housing and social grants provide a route out of poverty for the street homeless?

South Africa is not well prepared for increasing homelessness. Compared with the situation of people in shack settlements, little is known about the street homeless, and because they are so elusive only a few studies have been carried out (for example, Co-ordinated Action with Street People [CASP], 2000; Olufemi, 2002). There are no formal statistics. Although few governmental programmes address them directly, street homeless people are South Africa's poorest social sector: they are arguably worse off than the people in the shacks, many of whom have access to regular work, and to shelter, even if inadequate (Cross, 2008a,b). The nation's shack population is generally not destitute. The truly homeless are worse off: in their extreme poverty, isolation and loss of societal resources, they exactly fit the description ‘the destitute’.

Besides the consequences this severe poverty has for the nation's human rights undertakings, there is a tension between the national priorities of attracting investment in the cities on the one hand, and achieving poverty reduction on the other (City of Johannesburg, 2002, 2004). The effect of visible street homelessness on the prospects of economic investment in the metro core zones is understood by policy-makers as uniformly negative. The resulting clash between the rights and needs of the urban homeless poor to access street livelihoods in the central business district where they concentrate and the cities' potentially competing demand for a poverty-free central business district to encourage investment has made it difficult for government to develop a consistent homelessness policy. Together with the lack of basic information, this conflict has led to societal paralysis when it comes to solving the homelessness problem.

This article is a comparative introduction to the origins, extent and dynamics of the street homelessness problem in South Africa, seen in its world context. Section 2 considers the effectiveness of policies on homelessness, Section 3 examines definitions of homelessness, Section 4 looks at South Africa's urban homeless situation, and Section 5 notes homelessness trends in the industrialised and industrialising countries, including Britain, Europe, India and the US. Section 6 gives a comparative view of the historical development of homelessness trends in South Africa, and Section 7 contrasts the situation of the street homeless with that of the population living in informal settlements. Section 8 concludes by considering policy options and the way forward.

2. Policies on homelessness

South Africa has a deep belief in the capacity of society to perfect itself. Poor people are provided with free housing and infrastructure as an asset base to help those disadvantaged by apartheid to climb out of poverty (Hirsch, 2005). However, in spite of significant success with other social sectors and increasingly holistic and human-value-driven housing policy (Department of Housing, 2004), the destitute homeless are not finding this route.

Worldwide, homelessness is often seen as either a housing or a poverty problem (Tshitereke, 2009), and in South Africa as something susceptible to a solution through national housing delivery and social grants. The Human Sciences Research Council (HSRC) results show that government grants reach some homeless, which probably saves a good number of them from ending up on the streets, but so far there is no clear sign that free housing and infrastructure are reducing the numbers of street people already homeless: the HSRC's interviews with municipal and national officials and with caregivers suggest that very little subsidy housing is reaching this constituency.

Perhaps more than housing delivery, eliminating poverty may appear to be the obvious solution to homelessness. This is the position taken by many analysts, including several of the authors of the HSRC homelessness studies (for example, Ward & Seager, this issue). Others prefer to point to unaffordable housing or housing shortages (du Toit, this issue) or lack of institutional care, or all three, suggesting that street homelessness would almost disappear if these societal problems were dealt with.

However, people probably do not join the street homeless population only because they have low incomes and cannot afford their own housing or find a sheltering institution, although these are powerful contributory factors (see CASP, 2000). In South Africa, if formal subsidy housing is not being accessed, self-build informal housing should offer a cheap and generally available alternative. Besides asking whether poverty is the only cause, we also need to ask why the street homeless are not living in shacks.

It is vital to note that few, if any, nations have succeeded in eliminating homelessness, regardless of the size and reach of their social safety net. Homelessness remains a problem in the European Union, Canada, Japan and Australia – all highly developed countries that provide wide-ranging social care and work strenuously to eradicate poverty. It seems clear that poverty alone is not the problem. Likewise, it is not certain that South Africa, facing much more serious poverty than the industrial democracies, will be able to achieve what the world's richest countries have not.

3. Defining homelessness

After consulting stakeholders, the HSRC adopted for its study a definition of homelessness that emphasises living ‘on the street’. The HSRC qualitative interviews and quantitative survey work (see Kok et al., this issue) addressed adults and children who live on the streets full-time, in true homelessness or ‘rooflessness’ – the condition of routinely sleeping on the streets without regular access to shelter. However, UN-Habitat and other United Nations agencies emphasise inadequate shelter alongside lack of shelter as an aspect of the homeless condition, bringing the world's vast informal settlements into the homelessness arena (UN Centre for Human Settlements, 1990). Others go upstream to the pre-homeless condition, to include people in insecure or shared accommodation who are at risk of homelessness but not actually excluded from shelter as yet (Chung, 1991) and people who have become chronically isolated from society and social networks (Caplow et al., 1968).

These broader definitions risk losing the dynamics of actual street homelessness inside large and diffuse social categories: the destitute street homeless are a small and particular group as compared with the much larger, better known and less poor shack population and the many shelter-insecure urban dwellers. In South Africa, most poor people who do not have formal housing still do not become street homeless.

Little is known in concrete terms about South Africa's homeless street population, and there are no reliable estimates of their numbers. The 2001 Census estimates are regarded as flawed and unreliable. The HSRC's results suggest there may be from 100 000 to 200 000 truly homeless street people in South Africa's urban and rural districts together, including both adults and children. Johannesburg has the largest and most differentiated population – the HSRC's estimate for the city's street children alone suggested somewhat more than 3000. This figure tallies closely with the earlier estimate of Johannesburg street children by the Johannesburg Alliance for Street Children (Stone, 2004). This is a significant population, both in absolute numbers of the severely immiserated and in relation to the assumed economic impact of street homelessness at the local level on the goals of the developmental state.

Previous studies note that not all people on the streets are actually homeless in the sense of having no shelter of their own (CASP, 2000; Aliber et al., 2004). Street livelihoods can be surprisingly effective, pulling in people from surrounding settlements (CASP, 2000). These include piecework for local businesses, begging, foraging activity and sub-survivalist informal sector work, requiring business activity and a moneyed passing clientele. A significant population of ‘day strollers’ – with homes in shack settlements or townships – comes onto the street during the day seeking these livelihoods. The HSRC pilot study by Aliber et al. (2004) also identified street people who live on the city streets temporarily for street trading, commuting home to distant rural settlements. These non-homeless categories represent a significant share of the visible street population.

4. Homelessness in the cities

In practice, systematic policy addressing homelessness has tended to be paralysed by the lack of consensus as to what the effective options for intervention would be, and by the lack of clear information on the homeless population itself: homelessness largely represents a gap in formal policy (Naidoo, this issue). But the questions prompted by the homelessness problem test our key assumptions about the development process in the cities.

Although it is central government that sets social policy, much of the action taken to deal with homelessness is driven by empowered actors in local society – and actualised by the fears of businesses and local administrators – rather than coming down from higher levels of government urging the municipalities to act. Local government is often captive to the concerns expressed by business: the HSRC's interviews with local administrators underline the point made by CASP in 2000, that South African cities have tended to take uncompromising positions on suppressing visible street homelessness. Likewise, as interviews conducted for the HSRC study suggest, when either central or local government attempts more accommodating approaches, their power to put them into practice may be restricted by the lack of a clear policy consensus and the absence of data on the homeless population. The HSRC study attempts to address these shortfalls.

To promote the interests of the majority of their constituencies, cities aim to maximise their revenue base and promote high-end economic activity – since without a strong tax base, delivery will inevitably falter (see City of Johannesburg, 2002). While they may clearly see the need to accommodate the disadvantaged who can be drawn into productive activity, city administrators resist making systematic concessions to the need of the destitute for access to key urban spaces in order to support themselves outside the formal economy. However, as the cycle of micro-clearances repeats, the homeless themselves do not disappear but are pushed further into marginality – spatial, social and economic. The right of the poor to the city is not yet defined for the homeless.

5. The international situation

The kind of poverty that drives an unhoused poor population to move from place to place has been a major social problem requiring a policy response from very early times. In particular, vagrancy and begging have been seen as offences against society, but also as phenomena that society had some responsibility to prevent. Historically, the risk to social stability obliged governments to take an interest in the homeless and their welfare, in the hope of nipping in the bud mobile and dissatisfied social movements that might promote disorder.

5.1 Homelessness in Britain and Europe

The development of homeless policy in Britain is particularly important, since British colonial administrative practices spread throughout the world. The British vacillated between supportive and punitive approaches. During the Industrial Revolution, the numbers of unemployed people on the move increased as the rural population started to move to the urban centres. By the early eighteenth century, parishes in London had begun to impose laws on vagrants and beggars (London Journal 1729 cited in Norton, 2007). Mentally ill and disabled people were permitted to beg, but able-bodied persons caught begging were brought before the court and sentenced to hard labour or sent to a workhouse where they received shelter, clothing and food in exchange for hard labour (Cannon, 2001). To prevent employable individuals living at government expense, government relief was available to the destitute only if they accepted severely inferior workhouse conditions.

During the early twentieth century, the approach to poverty and homelessness turned to prevention and alleviation. Universal social insurance was available by 1946, and this was followed by the welfare state in 1948, which offered social services, old age pensions and unemployment compensation, official intervention to deal with low wages, and other benefits, in addition to free compulsory education which was already in place (Townson, 2001). However, the homelessness problem did not entirely disappear, and ‘rough sleeping’ continues today under the welfare state.

The twentieth-century wave of welfare reform was led by Sweden, which introduced family allowances, healthcare and other benefits in the late 1930s (Townson, 2001). Social benefits were expanded beyond a safety net to a very comprehensive public presence in the lives of citizens under the European Union, but homelessness persists in Europe at a fairly low level.

5.2 Homelessness in India

India, an industrialising state where rapid urbanisation involves large-scale displacements, offers important comparisons with Europe. Indian welfare policies maintain a socialist orientation, giving the government broad responsibility for anti-poverty work, while civil society fights for recognition of the complex and long-established street population (Huchzermeyer & Karam, 2006). As in Africa, rural districts sometimes maintain their own established urban informal localities where their migrants settle, keeping up their home ties while working on their own personal accumulation and economic advancement. Official toleration of street society is not consistent and clearances are common, but rural to urban migration continues to flow into a well-established informal economy that supports many homeless people outside the orbit of formal housing and control measures (Eyre, 1990). Indian authorities, unlike those in Europe, have limited control over the use of space by the poor on the streets, and street children without family contact are a major social concern.

Living options for people on the streets include the street itself, as well as squatting, homeless camps and slum-quality rental options (Glasser, 1994). However, attempts to reduce or eliminate street homelessness by providing urban public housing have often been unsuccessful because they peripheralise the poor away from their livelihoods – a problem that repeats across the homeless world, and resonates in South Africa particularly (cf. Huchzermeyer & Karam, 2006).

5.3 Homelessness in the US

The US took a different approach, downplaying the role of the state and emphasising individual responsibility and freedom in the face of poverty, homelessness and unemployment. Ideology in the US tends not to endorse government responsibility to provide livelihoods other than in the case of the disabled, and has been slow to support a government role in housing provision.

After the homelessness and widespread hunger of the Great Depression, public social provision expanded in the 1930s (Ware, 2004), partly in response to the problem of wandering tramps and hoboes. However, although slum clearance featured, large-scale public housing was not a focus, and a safety net for the poor was only put in place in the 1960s. Although states and cities provide welfare support, comprehensive lifelong welfare provisions on the European model create resistance and have never been attempted. Many US welfare programmes are highly conditional in an attempt to prevent free-riding, and homeless Americans have struggled to access government assistance (Regional Task Force on the Homeless [RTFH], 1999).

The US approach to homelessness emphasises affordable housing for homeless families and de-institutionalisation (Burleson, 2002), and the US is likely to differ from other developed countries in the scale of street homelessness among recently employed people with families. Thanks to the limited safety net, the extent of street homelessness closely tracks the state of the economy. Meanwhile, the homeless try to establish camps on the streets with shack-type structures, but are blocked by policing (RTFH, 1999).

Wolch and Dear's (1993) analysis of homelessness economics in American cities underlines the spatial factors that curtail access to work for the homeless and cripple their efforts to become reintegrated into society. As in other parts of the world, housing units torn down for high-end redevelopment are frequently not replaced, which creates a severe squeeze at the lower end of the housing market for working families. Displaced residents are forced to move to cheaper, less accessible areas further from work opportunities. Transport costs combined with low wages can limit or prevent access to available employment, resulting in serious economic marginalisation and increasing the numbers of the homeless.

This spatial mismatch between housing location and livelihood needs undercuts efforts to provide viable public housing for the homeless: suitable public housing spaces are rarely available in the high-demand city areas where the homeless find their livelihoods.

5.3.1 Learning from California?

In the face of limited help from the federal government, certain US municipalities adopted a bottom-up approach. San Diego's 25-year-old RTFH (1999) saw homelessness as a regional problem needing integrated responses at all tiers, and also developed a targeted care programme to bring back downtown investment.

This stakeholder-driven programme identified numerous regional homeless concentrations with different profiles, some with potential to move through transitional housing back into the housed society. But in addition to these outlying homeless, the programme identified a core downtown group of about 150 resistant, ‘hard core’ homeless: single men with behavioural problems, who had been on the streets for a long time and rejected outreach efforts. Recognising that there are different kinds of homeless people, a single one-stop-service downtown shelter allowed these entrenched long-term homeless to continue with street livelihoods, while concentrating on offering greater assistance to the less entrenched homeless who might more easily be reintegrated into society.

San Diego's programme offered assisted access to government support programmes to overcome bureaucratic complexities, and fed through to a transitional housing programme that emphasised very cheap, single-room-occupancy hotels on the streets. The Task Force acknowledged that the hard-core homeless had no homes to go to, and were unlikely to make the transition to independent living away from the streets. The programme therefore gave less priority to affordable housing than to emergency shelters and daytime facilities, and never treated moving or re-housing the street homeless as the first concern. San Diego's recognition that the street homeless population differs across the city region and its decision to manage city street homelessness in place rather than try to use housing to abolish it could both have implications for South Africa.

5.4 Trends in homelessness worldwide

Several of South Africa's homelessness issues appear in the world literature as the following pervasive themes or debates.

• The vagrancy theme identifies the homeless as an antisocial sector of society, a category of people thought to deliberately reject any constructive role, who can be viewed as social enemies. This viewpoint justifies action against or neglect of the homeless by the propertied classes. This theme can be traced partly to fears of the potentially violent threat a free-roaming, desperate category of people with no stake in society poses to governments, economic growth and property rights. When the homeless population is small, governments tend to ignore the problem and regard the street homeless as unimportant outsiders without citizenship rights, who can and should be excluded.

• The opposing social exclusion theme comprises the rights-based belief that providing social-democratic welfare will rehabilitate the street homeless population and thereby eliminate it. This view assumes that the homeless are not responsible for their marginality, which is a poverty and housing problem that can be cured by income measures and remedial housing. The nominal availability of so much help may baffle policy-makers and obscure the reasons why a street homeless population persists, remaining resistant to these remedial measures.

• Another theme is spatial displacement, urban peripheralisation and the search for livelihoods. This focuses on the street homeless population's need to have access to the central streets as their source of income. It considers the tension between the desire of the business community and the municipality to banish the street homeless from the metro core economic zone and the need to allow self-supporting street livelihoods – a tension that underpins the fruitless micro-clearances that serve for homelessness policy in major cities around the world.

• Last, and critical for South Africa, is the theme of how to understand the homeless street population in relation to the shack population. There is widespread uncertainty in the developing world, including South Africa and India, over how to address poverty and its effects on the streets and in the shacks, while the developed countries have been largely successful in eliminating shack settlements without being fully able to eliminate street homelessness. In South Africa, the size of the expanding shack population has attracted the most systematic policy attention and sidelined the question of effective homelessness measures, while the question of peripheralising the homeless is tied to the much more vehement debates over the peripheralisation of the shack population (see Huchzermeyer, 2005; Huchzermeyer & Karam, 2006).

In Europe's history, measures began to be taken against homelessness when the Industrial Revolution displaced much of the rural population from the land without supplying enough industrial jobs – a situation that is echoed in the industrialising world today. Social tensions over too few jobs leading to loss of shelter for the jobless fed fears of violence that persisted almost to the twentieth century, when the UK government began to provide shelter and regular relief. By mid-century, most working-class families lived in government-provided rental housing, and long-term unemployment on government welfare benefits had become common.

Today many street homeless people in Europe are able to claim social benefits if they do not have work, and are not for the most part forced onto the streets by unemployment or poverty alone. Most of those who remain on the European streets may be there because of social alienation, or as a personal protest against entering shelters. Others fall into chronic social isolation, having lost their social network resources after personal misfortunes. In itself, the world distribution of homelessness relative to safety net provisions in industrial and industrialising countries seems to rule out poverty as the simple, single cause of homelessness. Whether shortage of housing can be seen as the simple, single cause is another matter.

Compared with the industrialised European democracies, the US provides little access to government housing or stipends. The market cost of housing is high relative to working-class and middle-class incomes, and reportedly tends to leave many households and individuals displaced and stranded, unable to obtain work or afford shelter. The US government sees affordable housing as the main solution, while downward social mobility and falling through a weak safety net appears as a major cause of street homelessness. This is a risk that also affects South Africa's townships and suburbs, but here the dynamics are not identical to those found in either Europe or the US.

To overgeneralise, on the basis of the reported experience of the homeless in Britain, Europe and the US, the factor of social alienation is prominent throughout the industrialised West, being perhaps the main cause of homelessness that remains after the social safety net has alleviated immediate economic distress. This cause is less evident in South Africa, although alienation clearly affects the resistant homeless group. The economic causes of street homelessness bulk larger in America than in Europe, relating to different levels of societal commitment to assist those unable to live on their own earning power – a difference that reflects a deep transatlantic divide in values that define the nature and role of society and the individual. As Hirsch (2005) notes, South Africa broadly shares the European social values, but is at present unable to afford an equally comprehensive safety net for its much larger poor population. Although South African vulnerability to homelessness is narrowed by social spending, chronically unemployed men and older children may still fall through the safety net, and subsidy housing appears not to reach the street homeless.

The situation in India and other industrialising countries with a large poor population differs from Europe's, although it overlaps. Urbanisation, touched off by urban industrialisation, rural poverty and population growth, is leading to large urban-directed population shifts comparable with those of South Africa. For the destitute homeless in India and elsewhere, self-help informal solutions for both housing and livelihoods prevail, while very large street populations receive some official toleration. Such approaches raise issues of how far, given improved management, South Africa's metro urban centres might be able to tolerate a street homeless presence of the kind sustained in San Diego. However, South Africa's historical context appears significantly different from that of either Europe or the US.

6. Homelessness in South Africa

In South Africa, the street homeless problem affects both those who have fallen out of developed society and those who have never gained an initial foothold. Through most of South Africa's history up to independence, what was then called ‘vagrancy’ and ‘squatting’ represented inflamed and contentious issues that arose when landless black or coloured people moved around the countryside or tried to occupy unused land (Davenport & Hunt, 1974; Beavon, 1982). Under colonial rule, the white population sought to force the indigenous black population to live in designated areas so as to provide whites with a workforce, while preventing the emergence of an excluded and unsettled, migratory black population feared as tramps or vagrants, and as potentially an organised threat. As the dispossessed resisted apartheid efforts to compel compliance, the consequences were further removals and suffering among the displaced. The outcomes, in terms of street homelessness, are not well known: no statistics were ever kept on ‘vagrants’ or on the floating population without shelter.

The extensive livestock systems practised in South Africa led to acute conflict over land resources as the very large white farms claimed more and more grazing land, which squeezed livelihoods and severely undercut the economic autonomy of the black rural family. So-called ‘squatting’ – occupying land informally, without legal permission – became a prevailing problem of the colonies.

As landlessness increased from the mid-nineteenth century, large numbers of black people became labour tenants on white farms, and landless black families moved around to occupy any unused, formally white-owned land, as so-called squatters (Davenport & Hunt, 1974). Pressure on farmland was simultaneously giving rise to a new class of landless white squatters, who often succeeded in becoming bywoners (share-croppers or tenant farmers), competing with black people for available tenant-farmer places on larger white-owned farms, or otherwise moved into town. This homeless, mobile poor-white group also put pressure on the colonial government to remove black squatters from farms to make places for them.

By 1855 the Natal Colony had empowered magistrates to remove squatters from public or private land. The other colonies followed, with the result that there were wide-scale, continuing displacements and increasing population mobility, as well as loss of livestock and deepening poverty (Davenport & Hunt, 1974).

By the early twentieth century, the problem of squatting and displacement was enormous. Davenport and Hunt (1974:33) observe that ‘There were many more squatters on mainly white-owned land than there were inhabitants of the Reserves in Natal, the Orange Free State and the Transvaal’. These informal occupiers were chronically at risk of expulsion. Continual dismissals and evictions of tenants and squatters left numbers of black people adrift, as families or individuals, without land rights or community membership anywhere.

Homelessness occurred as the displacements increased, but – with the exception of the mobile ‘vagrant’ category – it was still largely absorbed within the black communities (Krige, 1962). The homeless rural ‘vagrants’ could work voluntarily for white farmers, who would provide precarious board and lodging in return for labour. Failing either, homeless rural-born individuals without close family to rejoin wandered the roads in white-controlled areas, looking for temporary jobs or domestic service work. This population was seen as a crime risk (Davenport & Hunt, 1974). From the date of the Cape's Vagrancy and Squatting Act (1878), legislation was passed in all the colonies to try to force the wandering homeless ‘vagrants’ into resident labourer status.

As numbers increased, this homeless rural population gravitated to the towns in search of work to substitute for the land-based livelihoods to which they no longer had access. By 1935 a report on the ‘Question of Residence of Natives in Urban Areas’ noted the presence of a floating population alongside the relatively stable and settled black population in all the country towns of the Transvaal (Young-Barrett Committee quoted in Davenport & Hunt, 1974).

6.1 The formal townships and the floating population

Although some formal black townships were established from the 1840s, urbanising black people mostly lived in ‘native locations’ on rent tenure from the municipalities under appalling conditions in temporary housing built of discarded materials (Union Government, 1914). The 1920 Housing Act provided municipalities with funding for housing the poor, but under the Slums Act of 1934 the integrated central cities were segregated through demolitions, and the displaced black population was removed to municipal housing ‘locations’ on the peripheries (Parnell, 1988). However, although always in danger of removal, the urban black floating population was not reported homeless in the sense of being completely without shelter, owing to the availability of the option of shack housing and the ease of obtaining sites before the advent of formal apartheid.

After 1937, apartheid spatial policies led to vast numbers of removals and dispossessions in both the urban and rural sectors. While most of the dispossessed went on to other townships or to shack settlements and visible homelessness was not permitted on the city streets (Davenport & Hunt, 1974), there is no way of knowing how many removed people may have become itinerant ‘vagrants’ or street homeless, making a precarious living in rundown city areas or abandoned buildings, with no permanent shelter.

Pressure on available township housing increased again in the 1960s, and by the 1970s, with government housing delivery diverted from the cities to the rural former black reserves, the housing waiting list for approved urban residents in Johannesburg had reached 11 000 (Surplus People's Project, 1983). It appears likely that by the 1970s and 1980s people unable to find their own housing had started to flow into the now-prohibited shack settlements around the towns and cities. Evictions from farms were particularly significant in causing displacements (Surplus People's Project, 1983).

The size and location of this floating population, the dislocated and alienated urban homeless sector, at that time caused limited concern compared with the mushrooming shack population. However, it appears that the street homeless population in the central cities has emerged into the open since the deterioration of apartheid policing control over urban space ahead of the democracy elections of 1994 (Cross et al., 1992). Other than general rights commitments on shelter, no clear policy has yet been framed (see Naidoo, this issue), but the historical British fear and distrust of homelessness have become South Africa's legacy.

6.2 South African street homelessness today

At present, the street homeless population is clearly separated from the shack population living in informal shelter. Homelessness affects people of all races in South Africa today, although the majority black population makes up the largest component. Homelessness is found today both in the urban poverty sector and in the rural districts, and appears to be functionally descended from South Africa's displaced rural ‘vagrant’ population of the nineteenth century.

For the new South Africa, little has yet been written about the street people, and virtually nothing about homelessness in rural areas. Only a few research studies have been carried out on the street homeless adult population in the metro centres. More work has been done on the phenomenon of street children (see Ward & Seager, this issue), which is often seen by policy-makers as a more urgent priority. However, taken together, these pioneering studies go far to describe the situation of the street homeless population in South Africa today.

For the adult population, the best known study is by Olufemi (2000, 2002), who pioneered work on homeless women in Gauteng. This study highlighted the differences between men's and women's homeless experience, noting that homeless women were significantly fewer than homeless men.

The Cape Metro Council Street Field Workers Project (CASP, 2000) carried out an audit of street people in the Cape Town area. CASP was able to show that there are many more adults in the metro street population than children, although adults receive much less attention. A more recent study by the Street Alliance network of non-governmental organisations (Stone, 2004) also noted that the population of street children in Gauteng is very visible but relatively small, which opens up possibilities for intervention.

The overall picture that emerges from the South African research literature is one of a small and sometimes severely alienated group comprising mainly male adults, but also including women and children. Inside the severely alienated group, the CASP study identifies a category of resisting street homeless who reject shelter and civil society assistance programmes, analogous to the hard-core homeless of downtown San Diego.

Although South Africa has developed little specific homelessness policy, in trying to address the consequences of apartheid it surpasses almost all other developing countries in making available anti-poverty programmes, including free subsidy housing, comprehensive human services and direct income support. As with the case of Europe, it seems vital to ask why South Africa's well-resourced and well-targeted national anti-poverty initiative has not yet eliminated the problem of homelessness from the country's streets.

7. Homelessness and informal housing

South Africa is both different from and similar to the industrialised and industrialising countries; specifically, the presence here of a large shack-resident population among the poor raises questions about how street homelessness relates to housing provision.

In this light, it is important for understanding South Africa's street homeless population to decide whether it is defined together with or as distinct from the population of the shack settlements. The two groups do not necessarily behave in the same way when it comes to livelihoods or housing options, and in many ways they are in opposite situations.

The street homeless in the US and Europe are not necessarily the poor, and often appear to come from the middle class: the cause of their homelessness is often economic or social displacement and downward mobility rather than pre-existing poverty. However, it is key that in the US and Europe almost all housing is formal: self-build informal housing is efficiently prohibited, specifically including attempts to build on-street homeless camps for living accommodation inside the cities. Therefore, apart from some squatting in abandoned buildings, the industrialised world has no real equivalent of shack settlements as an immediate shelter option for the marginalised that is under their own control without reference to bureaucracy.

In most of the industrialising world, informal self-build shelter for the poor in cities is the norm, and South Africa is no exception (Rogerson, 2000). However, the street homeless in South Africa are unable to make significant use of the self-build shelter option, owing to the locational mismatch that separates the central spaces where they make a living and the less-defended peripheral areas where informal housing can still be built. The few established inner city shack areas are in high demand from work seekers, tightly packed and relatively expensive, and are prevented from expanding: the street homeless need not apply. At the same time, South Africa's street homeless are rarely able to penetrate bureaucratic requirements to obtain subsidy housing, while subsidy housing is similarly located out of reach of the spaces they need. Like the American homeless – and in spite of high public spending – even those South African homeless who would be open to housing solutions may be left stranded on the streets without effective access to either formal or informal housing opportunities.

7.1 The shack connection: Differentiating homelessness

Meanwhile, most rural–urban migrants in India and South Africa do not go onto the streets but rather into the shack settlements, mainly on the urban periphery. The street homeless here appear as a separate population: homeless street people concentrated in the city centres appear to be predominantly unskilled and often demoralised single men and older male youth who have lost contact with families and have little chance of employment (see Kok et al., this issue). Many may belong in the ‘resistant’ category that rejects most or all assistance.

Often ambitious and actively work-seeking, the metro shack areas population appears better connected, often comprised of small urban families or migrant workers who have families elsewhere and a community identity, as well as single work-seekers and others. As many as two-fifths of adults here report holding formal jobs (Cross, 2008b), which allows them to afford informal settlement costs and the necessary transport; and shack-resident women often survive on child grants rarely accessible to the isolated street homeless. Unlike the street homeless, the people of the shack settlements participate in the normal society of poverty in South Africa. This apparent filtering of the rural-to-urban migration stream into different social sectors with different apparent levels of capacity and connectivity brings our attention back to the issue of how to define homelessness.

Results from recent HSRC/Council for Scientific and Industrial Research work (cf. Cross, 2008a,b) suggest that the shack population is less severely excluded than the street homeless, and may already have some of the advantages of home ownership that the government housing policy (see Hirsch, 2005) strives to promote. That is, as in India, the shack residents may use their informal housing with access to the city to support some savings and accumulation activity (Cross et al., 2005). If so, in order for accumulation to take place, shelter is probably necessary as a stable platform for household operation, but it may not always need to be good quality housing with title as long as it is reasonably secure in practice and allows for some future planning.

In contrast, the street homeless population remains adrift with little if any means of savings or accumulation. Although street livelihoods yield enough to draw in income-seekers from the shacks and from rural areas, without some kind of housing as a platform it appears that living full-time on the streets may be mainly hand-to-mouth, with little option of future goals: worldwide, the homeless concept of time is often circular and repeating rather than linear and goal directed (Glasser, 1994). In these respects, developmental dynamics and appropriate policy options for the shack constituency and for the street homeless may almost be diametrically opposed, so that conflating the situations of these different social sectors may be obscurative and counterproductive.

8. Some conclusions on homelessness and housing

South Africa's street homeless are a small and highly specific group compared with the shack population. They differ in access to employment and to government benefits, and a resisting core group has been identified that commonly rejects attempts at help. In their constraints, the street homeless here may not closely resemble the homeless populations of either the industrialised West or industrialising nations such as India. Whether the shelter options developed for this constituency will include informal self-build housing or better access to existing shack areas – in addition to formal shelters, cheap rental options, homeless people's drop-in centres, and help with transport and formal housing delivery – remains an open question.

Government housing policy is surely correct in identifying access to housing as critical to escaping poverty through household accumulation. However, housing policy is targeted at the shack population and not at the less capacitated homeless, and there may be reason to doubt anyway that household accumulation depends only on the delivery of good quality housing for this poor constituency. Going on California's experience and on civil society advice, independent permanent housing – however affordable – is probably not an option for many of the South African street homeless.

If South Africa has a resistant street population unwilling to accept either shelter or housing programmes, their situation will need to be managed while they remain on or near the streets. If so, total abolition of the homeless condition may not be a realistic option under any policy dispensation, and management then becomes the key concern.

Shack housing is the other alternative. Sometimes described by respondents as being just as good as the lower grades of subsidy housing (Cross, 2004; Kitchin & Ovens, 2008), shacks may often meet some if not all of the requirements for catalysing self-savings behaviour as a route out of poverty. If so, then existing shack housing is a rung on the accumulation ladder for those who have it, and may be an option for anyone – homeless or work-seeking – who needs accommodation with location advantage.

It would appear that neither route is normally open to the street homeless, who lack the usual tools of accumulation in the form of secure housing and incomes. South Africa follows the western democracies in excluding new self-build informal dwellings from built-up city areas, and residence in civil society homeless shelters does not appear to start a future-directed accumulation process. Nor do the street homeless report successful relations with administrative offices: although entry to a shelter would provide an address and most have tried shelters in the past, most respondents said they had no identity documents and few had any ongoing relation with government programmes. Very few said they had grants or were on housing lists. Without more social workers specifically tasked with outreach and support for the street homeless in dealing with administrative offices, it is not easy to see how the street constituency would be able to become regular beneficiaries of subsidy housing, and the requirement of a residential address is a barrier for those street homeless who reject shelters. Consequently, chances are slim that the homeless will start successful savings and accumulation based on their street livelihood earnings.

At the same time, the cities' administrative stance tends to remain distant or hostile, relying largely on police clearances to deal with the resistant sector of the homeless population who are unwilling to accept civil society help and enter shelters. Criminalisation of the street homeless emerges from concerns of the business community and other economically committed interest groups who perceive the need to avoid having poverty evident on the city streets. Criminalisation is perhaps mainly due to the absence of any widely recognised and effective remediation options that such interest groups can adopt and support, and it spreads from the contested central zone outward.

If this is the case, criminalising the homeless might be addressed by an education initiative combined with San Diego's type of differentiated, multilateral city-region approach, which stresses management by amelioration in place of attempting to eliminate or banish street homelessness. Offered a wide range of optional services without compulsion, San Diego's resistant street population has not been eradicated and maintains its right to the central streets, but may become by degrees more comfortable, more healthy, and also perhaps less alienated, more engaged and more presentable in the eyes of other stakeholders.

Overall, it seems clear that the homelessness question is not limited to shelter, and is not a simple matter of poverty either. For the developmental state, the street homeless are the proverbial skeletons at the feast, the excluded poorest who enter unobserved and stand by gaunt and starved, terrifying to the invited guests but deprived of any capacity to join the party. For the shack residents – whose poverty situation is well recognised – South Africa has consistently worked to provide a place at the development table, but for the street homeless the anti-poverty entitlements offered by government remain on the table in front of them just out of reach. To bring the homeless to where they have scope for accumulation and rehabilitation without by their presence undercutting the cities' chances of economic expansion, the key may not be suitable housing in itself, but livelihoods and access to central urban spaces.

Acknowledgements

This work forms part of the Human Sciences Research Council's 2005–2008 study of homelessness. Funding from the National Department of Social Development, the HSRC, the Gauteng Department of Social Development and the Swiss Agency for Development and Cooperation is gratefully acknowledged.