South African street children: A survey and recommendations for services

Abstract

This paper describes a sub-study of children within the Human Sciences Research Council study of street people in South Africa. It used both quantitative and qualitative methods (a survey and census, and focus group discussions and in-depth interviews) to gather information from street children, their service providers, and five women who had grown up on the street but subsequently made a success of their lives. The paper reviews the origins of homelessness among children, the dynamics of life on the streets, and care requirements. It recommends a continuum of care for those already on the streets, but points also to the need for preventive services. It suggests that truancy from school could provide early warning of children at risk, and recommends a range of family support services that may be needed to prevent children from taking to the streets.

Keywords:

• street children
• homeless shelters
• childcare services
• South Africa

1. Introduction

Within the Human Sciences Research Council's broader study of homelessness, on which this special issue is based, the authors conducted a specific study of homeless children. The purpose of this was threefold: to examine the current status of South African street children, to estimate the number of street children in Gauteng (a figure that was needed by the Gauteng Provincial Government for service development), and to make recommendations for services.

‘Street children’ are likely to be more diverse in circumstances than homeless adults. Children may be working or begging on the streets while living at home, they may not return home every night but still have frequent contact with their families, they may have lost contact with their families entirely, or they may move between these conditions (Baker, 1999; Panter-Brick, 2002). Terms widely used to describe them are ‘children on the street’ – those who are visibly working or begging on the streets but continue to live with their families – and ‘children of the street’ – those who no longer have contact with their families and may be regarded as homeless (Ennew, 1996). Recent South African legislation, the Children's Act (Act No. 38 of 2005), does not distinguish between these two types but includes both as ‘children in need of care and protection’.

Our approach was to focus on the ‘absolute homeless’ or ‘roofless’; in other words, we excluded ‘children on the street’. However, we recognise that such children are at risk of becoming absolutely homeless – at risk of moving from being children on the streets to being children of the streets. This notion of risk is addressed in our recommendations.

2. Methodology

Three related methodologies were used – interviews, survey and census.

In Johannesburg, two shelter managers, three project managers, two fieldworkers and five care-workers from three programmes serving homeless children were interviewed. In Cape Town, a focus group discussion was held with five directors from four service providers, and one manager was interviewed individually. These interviews assessed the service system environment.

Focus groups were held with children and young people living in shelters and still on the streets, in both Johannesburg and Cape Town. Shelter outreach workers were used to contact children on the streets and to invite them to participate. Males and females were interviewed in separate groups because there are gendered aspects to going onto the streets (such as sexual abuse at home) that respondents might not have been comfortable discussing in mixed groups. Some respondents refused to be interviewed in a group and were interviewed individually. In total, 52 young people living in shelters and 46 not in shelters were interviewed. Children under 12 years old were excluded because it was considered too difficult to get informed consent from them; a few young adults who congregated with street children were included in order to get their views of the situation of children on the streets.

A mini-census was conducted of street children at their points of interaction with services for the homeless, such as ‘soup kitchens’, or their sleeping places. This was used to estimate the number of children of the streets in Johannesburg. Subsequently, a survey was carried out, using semi-structured interviews. Shelter outreach workers, who were trusted by the children, helped to recruit participants. In total, 305 children between 12 and 17 years of age were interviewed.

In Cape Town, five women who had been served by a street children's programme as children, and had achieved stable adult lives, were purposively sampled. The choice of women was fortuitous and happened because we were working closely with a girls' shelter. (This is a limitation of this study – the inclusion of men would have strengthened our data.) These interviews explored factors that contributed to achieving successful outcomes. The five women were all in stable employment, three had degrees or diplomas, four were married and three were raising their own children. Two of the women were coloured and three were black, four were in their late twenties or early thirties, and one was in her forties. They were invited to describe how they came to be on the streets and explain what had helped them achieve stable adult lives. Their stories overlapped in many ways with the results of the survey, and we therefore concentrate on reporting what is unique in their contributions.

Participants were interviewed in the language of their choice. Where necessary, the interviews were transcribed and translated into English, and participants' actual names were removed from the transcripts to preserve anonymity. Transcripts were analysed using template analysis (Crabtree & Miller, 1992), aided by the software package Atlas.ti. In template analysis, a ‘template’ of themes is developed from a subset of data and used to analyse all the data. The template is modified when new themes emerge, or prove to be rare in the data.

The study was approved by the Research Ethics Committee of the Human Sciences Research Council.

3. Results

3.1 Demographics of the survey sample

In order to estimate the number of street children in Gauteng, several assumptions were made. For example, it was assumed that they would be found in the vicinity of shelters, since these are most likely to be located where they can easily reach street children and the children would come to the area to use the services they offer.

An audit of the catchment areas of the shelters showed that, on average, there were 29 children in the catchment area of each of the 60 known shelters in Gauteng. On the basis of information gathered from audits of 10 shelters, it was estimated that nine out of 10 such audits would find fewer than 39 children in the catchment area of each shelter – or that there was a less than one in 10 chance that, had every shelter in Gauteng been audited, fewer than 2300 street children would have been identified. However, the first attempt to recruit child interviewees identified 157 children, of whom 27 per cent lived in areas that were not catchment areas for shelters. We could thus assume that there were children who did not live in shelter catchment areas, and revised our estimate accordingly. Doing so meant that there was a nine out of 10 chance that there were fewer than 3200 homeless children in Gauteng at the time of the survey (2007).

It must be emphasised that counting the homeless population accurately is an extremely difficult task and, because of the assumptions involved, 3200 must be regarded as an upper limit. This estimate is strikingly close to the figure of 3102 children identified on Gauteng's streets by the census conducted in 2005 by the Gauteng Alliance for Street Children (Roestenburg & Oliphant, 2005).

Table 1 presents demographic details for the sample. The percentages given in this and the remainder of the paper are sample proportions and not population estimates. The basis for our estimation is not strong enough to permit generalisation from the sample to the population.

Table 1: Demographic details of the 305 children interviewed

	Number	Percentage
Number of street children who said they are South African citizens	233	76.4
Number of male street children	284	93.1
Race of street children
Black	297	97.4
Coloured	6	2.0
Marital status
Married	1	0.3
Never married	301	98.7
Current sleeping arrangements
On street	230	75.4
In shelter	12	3.9
In park	28	9.2
With friends or associates	2	0.7
Other	33	10.8
If living on street
Has no permanent home	82	26.9
Is temporarily stranded	33	10.8
Lives too far away to commute	185	60.7
Other	4	1.3
Number of street children who themselves have children	9 (5 male, 4 female)	3.0
Number of children the street children have	6 reported 1 child, 1 reported 3 children

Regardless of whether these percentages over-estimate or under-estimate the population, they give some indication of the situation of children on the streets of Gauteng. One notable characteristic is the low number of girls. Some studies suggest that girls are less likely to become homeless, because they may more readily be absorbed into households, probably because they are perceived as being able to contribute to the household through doing housework and as being less troublesome (Swart-Kruger & Donald, 1994). Service providers said girls who became prostitutes were less likely to access shelter services because they had an income and, often, housing; this is echoed by the literature (Richter, 1991). So it is likely that there are fewer girls than boys among street children in Johannesburg. However, there may also be sampling bias; girls may be more wary about being approached by a stranger than boys and may therefore have kept away from the fieldworkers.

Figure 1 shows the ages of the boys and girls (sampling was restricted to children between the ages of 12 and 17). The average age of the children in our sample was 15.71 years.

Figure 1: Age and sex distribution of the 305 children in the sample

3.2 Reason for leaving home

In both the survey data and the qualitative data, children described ‘push’ and ‘pull’ factors that influenced their taking to the streets: they left to escape intolerable situations, and in the hope of finding better ones. Of course, the study has limitations – children were asked only one question in a survey, or were interviewed once by a stranger, and may have been reluctant to reveal the more painful reasons why they left home, such as sexual abuse, exploitation or abandonment (Schärf et al., 1986; Richter, 1991; Swart-Kruger & Donald, 1994; Baker, 1999). The response frequency analysis should therefore be treated with some caution.

In the survey, only 28 children (or 9.8 per cent) nominated a pull factor (seeking ‘a better life’, employment, money, shelter or survival) as their primary reason for going onto the streets. Three children (1.0 per cent) reported that they had gone onto the streets to search for one of their parents, with remarks such as ‘My mother deceased and I left home to search my father’.

For the most part, however, children described push factors. Many (54 children, or 18.9 per cent) described situations of abuse, domestic violence or poor family relationships that had made their home lives intolerable. In our more in-depth interviews with children, girls described sexual abuse, typically by stepfathers or their mothers' boyfriends, while boys described vexed relationships with step-parents. In many cases, they described these situations as being fuelled by alcohol. Six children (2.0 per cent) simply said they had run away, without giving any explicit reason for having done so. However, it is unlikely that children would have left a happy home; rather, ‘running away’ is likely to be a response to a family situation characterised by conflict or violence.

Loss of caregivers – because both caregivers had died, or because one parent had died and the whereabouts of the other was unknown, or because parents had left home to find work and lost contact with their family – was the reason given by the largest number of children (79 children, or 27.6 per cent). Estimates suggest that over 10 per cent of South African children have lost one or both parents from AIDS alone (Grassly et al., 2004), and migration from one area to another to find work is a common practice among the poor in South Africa (Aliber et al., 2004). It is therefore not surprising that these are the leading reasons supplied by children for leaving home. However, their answers here demonstrated that it was not simply the loss of a caregiver that led to their leaving, but their own vulnerability to abuse when their caregiver was no longer there to protect them.

Sometimes this vulnerability was to poverty rather than abuse – several children gave answers such as: ‘Both parents deceased so my family members cannot afford to support me’. Their answers also demonstrated failure of the extended family to act as a safety net: one child, for instance, said ‘I am an orphan, all my siblings are married and no one looks after me’. Having said that, however, a few children reported or implied that they had left home with their parents when their home had burned down or their caregiver was looking for a job. Shack fires in informal settlements are a common cause of homelessness among the urban poor (Harte et al., 2009).

Poverty was given directly as a reason for leaving home by 35 (12.2 per cent) of the children, while other answers also implied poverty, such as not being able to afford the rent (two children) or unemployment (two children). It was not possible to tell, in these cases, whether the respondents were referring to their own inability to pay rent or their own unemployment, or that of their caregivers.

Seven children (2.4 per cent) said they had left home because they had dropped out of school. Service providers in Cape Town emphasised that playing truant from school, or dropping out of school, is often an early sign that a child is likely to take to the streets. Again, this suggests the reasons are poverty – the parents' inability to pay school fees, or for uniforms or other things that help children feel included at school – and family systems that are unable to assist children who most need help. Other studies suggest high rates of learning disabilities in this population: children who have learning disabilities and do not receive the necessary support are likely to find the school environment punishing rather than pleasant, which increases the risk of their dropping out (Richter, 1996). This suggests a need to intervene in the school system (to find ways to help poorer children meet the financial obligations of school, and to provide the necessary support to those with special needs). It also suggests where efforts at prevention might profitably start. Children who are poor and who begin to play truant from school are easily identifiable through the school system, and early intervention before they drop out of school may help prevent them from taking to the streets.

Seven children (2.4 per cent) said they had left because they had committed a crime and feared retribution, while one nominated his own ‘bad behaviour’ as the reason. In the more in-depth interviews, children who admitted to difficult behaviour often also described a loss, such as the death of a parent, and their own grief as causes of their behaviour, but their stories revealed that they had received no support from caregivers in their grieving. It is likely that these children were depressed, a mental health problem that, in children, often manifests as oppositional, defiant and difficult behaviour (American Psychiatric Association, 1994). The qualitative data also showed a link between substance use and crime: some children admitted that they were using drugs and alcohol, and had been committing crimes in order to support these habits. These stories may be evidence of unidentified and untreated mental health and substance misuse problems among children that, if treated, might have prevented their taking to the streets. Although the rates of these disorders are not known nationally, the prevalence of depressive disorders among children under 18 years old in the Western Cape is estimated to be 8 per cent (Kleintjes et al., 2006). Rates of treatment, however, are not available.

Five children (1.7 per cent) were simply lost, and it appeared that they had come into the city with friends who promised them something (such as a party) and then abandoned them without the means to make their way home.

Fifteen children (5.2 per cent) nominated what we coded as ‘peer pressure’ as a factor in leaving home; that is, running away with friends, or being influenced by friends to run away. From the qualitative data, it seems that peers who were already living on the streets, or who knew someone who was, made them think that life in the big city would be easier and fun, and that it would provide access to material goods or the means to make a living.

Finally, 30 children (10.4 per cent) said they had left intolerable conditions (either because of civil war or the economic situation) in their home country.

While the survey provided a broad picture of many children's reasons for leaving home, it was inevitably a rather superficial glance at a complex issue. The narratives of the five adult women who were interviewed, however, do give more nuanced pictures. For instance, one woman described how she and her older sister had lived with their grandmother in the Eastern Cape until she was 12 years old, when her grandmother died. After being let down by relatives in the Eastern Cape, they had then decided that their best chance of a stable living situation was to travel to Cape Town to find their mother. Together they caught a bus to Cape Town and found her. However, unbeknownst to them, their mother had suffered brain damage in a robbery several years earlier, and now was prey to paranoid delusions. They spent one night with her but she threatened to attack them while they were sleeping. It was clear that their mother could not care for them, so they sought the care of an aunt; the aunt's husband, however, did not wish to assume the care of two teenage girls, and so our participant then sought out her father's relatives. She managed to find some of her paternal aunts. She did not reveal clearly what happened in that household, only stating that she had hoped to be welcomed there but found that she was not, and that because of her situation she was unable to go to school, which disappointed her. It was at this point that she gave up trying to find stable housing with relatives and took to the streets.

We have included one story in detail because in many ways it is typical of street children's stories. It shows that the step of taking to the streets is not taken lightly in most cases (if any), but is the result of a complex chain of events in which a child has demonstrated extraordinary tenacity and initiative in seeking to keep herself housed and to elicit appropriate care from adults who consistently fail her.

The reasons for leaving home reported in our study do not seem to have changed much since earlier studies of street children in South Africa (Schärf et al., 1986; Richter, 1991; Swart-Kruger & Donald, 1994; Baker, 1999). Together, the children's stories and the literature show that the presence of street children is a symptom of broader social problems. The backdrop to a child's choice to leave his or her home for the streets is widespread poverty, rapid industrialisation, urbanisation and westernisation that have weakened extended family systems, and hence the ability of families to provide a safety net for children (Swart-Kruger & Donald, 1994). Most recently, some authors suggest that the HIV/AIDS pandemic may exacerbate this, and hence increase the number of street children (Baker, 1999; Ward, 2007).

3.3 Time on the streets

In our survey, as in others (Donald et al., 1997), we found a very large variation in the time children had spent on the streets (see Figure 2). This is an important index for rehabilitation: service providers reported that, in general, children who have spent less time on the streets before entering the shelter programme are easier to place back with their families. In part, this is because the child has made a choice not to be on the streets; and in part it is because the less time a child spends on the streets, the more likely he or she is to have retained habits that will facilitate integration into mainstream society (such as getting up in time to go to school). In our survey, the average length of time that children had spent on the streets was one year and nine months, but there was a wide range within this group, from six days on the streets to 17 years. This last figure implies that some children in our sample had spent most of their lives on the streets.

Figure 2: Years children had lived on the streets

3.4 Migration patterns

Children may migrate considerable distances from their original homes. Our data suggest that this was true of many of our sample – only 128 children (42 per cent) from the Johannesburg survey sample had been born in Gauteng. Shelter staff reported that children often begin their lives on the street close to home, but drift towards city centres, where they are most likely to find opportunities for food and work. Another pattern identified by shelter staff, and confirmed in the literature, is that a disproportionate number of children are likely to have come from the same small area, sometimes as small as the same block (Cockburn, 1991). This indicates an area where children are at high risk of a range of problems, such as dropping out of school or being the victims of family violence, and that, because they know other children who have successfully taken to the streets, they may work up the courage to do so themselves.

Unfortunately, the fact that most street children end up in city centres also means that most of the services available to them are located in city centres and reach them only when they have been on the streets for some time. A more successful strategy might be to locate the services in the neighbourhoods where children start their journeys. This would improve the chance of intervening early and reuniting them with their families before they have lost the habits associated with living housed, or preventing them from leaving home in the first place.

3.5 Livelihood strategies

Children living on the street reported that their basic needs such as food and clothing were typically met by soup kitchens, church organisations, the Salvation Army and ‘ordinary people’. In terms of accommodation, children congregated in places such as specific street locations, parks, unoccupied buildings and a taxi rank. However, they also reported engaging in various activities to earn money.

Of the children interviewed, 30 (9.8 per cent) reported that they had a job or were self-employed. Incomes varied quite widely: 35 children (11.5 per cent) said they had no income, and the majority (229 children, or 75.1 per cent) said their income was under R200 a month. A further 5.3 per cent reported incomes varying from R400 to R1000 per month, and one child reported an income of between R1100 and R1500. As some contextualisation, income data for 2007 (when this survey was conducted) indicates that the mean per-capita income for the poorest 10 per cent of the population was R1032 (Republic of South Africa, 2008). All but one street child, therefore, reported earning well below this mean, suggesting that life on the street is certainly not life on Easy Street.

Aside from the children who were working, a further 119 children (39.0 per cent) were looking for work. They were, for the most part, looking for unskilled work, such as gardening, helping with bricklaying, domestic work, hawking or guarding cars. One-half of the sample was too young to work legally (under 16 years old) and most, therefore, rely on begging, serving as car guards, and washing cars.

Our survey did not cover illicit methods of earning an income, but the qualitative data showed that gambling and stealing were other strategies. While very few admitted to stealing, nearly all boys reported having been arrested at least once for theft and had served their terms or been released to parental care after being warned by magistrates. Some of the preliminary work completed in Durban suggested that street children are vulnerable to threats by gangs, who may force them to carry out petty theft for them.

Our work suggests, therefore, that while criminal activities are certainly among the livelihood strategies used by homeless young people, they are neither the primary nor the only ones. Nearly 70 per cent of our sample was either working or looking for work, and the bulk of the remainder was too young to work. It also appeared that criminal activities might be survival strategies, or might result from needing to support a substance abuse habit or from fear of established criminal groups. To resolve this issue will require a study over an extended time that can gain the trust of wary young people: a survey and focus groups were unlikely to provide such sensitive information.

The qualitative data also revealed that girls used different livelihood strategies from boys. While many reported gambling as a means of earning money, they said they primarily relied on boyfriends for material goods. They emphasised that if they did not have a boyfriend while on the street, they were vulnerable to rape by other men. In a study of neighbourhoods with high levels of gang activity, Ward and Bakhuis (2009) found that the young women reported that liaisons with men provided some protection, as well as access to material goods, and that the pressure to raise money for girlfriends might concomitantly play a role in driving the young men into criminal activities. The situation was found to be similar among the street children. However, prostitution was reported to be uncommon among homeless girls. Service providers in Cape Town concurred with our data in stating that prostitution was rare, although possibly because girls involved in prostitution had a means of livelihood and may have been housed by their pimp, they were not viewed by themselves or their communities as street children and were also unlikely to access the services offered to street children.

In the qualitative data, children reported that they used the money they earned mostly to buy drugs, food and clothes. The survey sample demonstrated that a high proportion of street children (37 per cent) shared their money with other people, implying that they were working in some form of network that may have had either protective or coercive aspects, or both.

3.6 Education and skills

Many street children report school failure, extreme punishment at school, or an inability to attend because their parents were unable to pay fees or to buy uniforms (Schärf et al., 1986; Richter, 1991; Swart-Kruger & Donald, 1994). Most are therefore functionally illiterate. While most will also say that they would like to return to school, few will find this unproblematic. The majority are likely to be handicapped by being too old for the grade to which they need to return, or by learning or other disabilities (Richter, 1991). In our survey sample, the majority had a Grade 7 education or less, and 175 children (57 per cent) were old for their grade (see Table 2).

Table 2: Highest level of education (completed grades) by age

	Age						Total
	12 years	13 years	14 years	15 years	16 years	17 years
No schooling	0	0	0	0	0	4	4
Sub A or Grade 1	0	0	1	0	0	0	1
Sub B or Grade 2	1	1	0	0	0	1	3
Grade 3 or Standard 1	1	0	3	1	2	3	10
Grade 4 or Standard 2	3	0	2	0	3	5	13
Grade 5 or Standard 3	6	4	4	6	7	5	32
Grade 6 or Standard 4	2	8	4	5	9	11	39
Grade 7 or Standard 5	1	2	6	10	11	21	51
Grade 8 or Standard 6 or Form 1	1	5	6	10	10	16	48
Grade 9 or Standard 7 or Form 2	0	0	3	6	9	21	39
Grade 10 or Standard 8 or Form 3	0	0	0	3	7	21	31
Grade 11 or Standard 9 or Form 4	0	0	0	1	4	23	28
Grade 12 or Standard 10 or Form 5 or Matric	0	0	0	0	1	4	5
Diploma or certificate with Grade 12 or Standard 10	0	0	0	0	0	1	1
Total	15	20	29	42	63	136	305
Boxed cells indicate normal age for grade; bold figures and shaded cells indicate over age for grade.

While many of these were not more than one or two years older than would be expected for their grade, it is highly likely that they will have had a long period of disrupted schooling prior to taking to the streets, and reintegration into mainstream schooling would not be straightforward. For several of these children, the problem was dramatic: for instance, four of our sample, aged 17, had had no schooling whatsoever. For young people in this predicament, traditional mainstream schooling is not the answer. Alternative education programmes that emphasise functional literacy and a marketable skill (as are provided by several shelters) are therefore essential for this group.

3.7 Access to welfare services

Apart from some patients at HIV/AIDS treatment centres who were reportedly receiving disability grants, access by homeless youth to welfare services such as social grants was problematic. Primarily this was because most homeless children do not have identity books.

This problem can seriously affect all street children's access to grants to which their caregivers are entitled, and is problematic both for young mothers who live on the streets (if they do not have identity documents, then their children will not have birth certificates, and they will be unable to access the child support grant) and for shelter staff, who wish to access grants for children's families to facilitate their return home. Grants can help to reunite a child with his or her family successfully, as the financial burden of an extra child on the family is relieved. In addition, the lack of an identity book has a serious impact on the shelter's ability to obtain health services, education or employment for a young person in their care. In our study, only 107 children (35.1 per cent) said they had either a South African identity document or a birth certificate.

3.8 Services to street children

In our survey, children were asked what they thought was ‘the most important thing the government can do to help’. The most frequent responses were provide employment (80 children, or 26.2 per cent) or housing (79 children, or 25.9 per cent). Most of those who suggested housing wanted a place in a shelter (51 children), followed by affordable self-owned housing (17 children) and affordable rental housing (nine children).

Other forms of assistance the children nominated included skills training (54 children, or 17.7 per cent), grants or money (23 children, or 7.5 per cent), primary healthcare (four children, or 1.3 per cent), and emergency assistance (two children, or 0.7 per cent). Other children nominated other forms of assistance, including ID books or residence permits (14 children), reuniting families (nine children), or assistance with returning to school (13 children).

Street children's programmes offer a range of services, such as outreach work, drop-in centres, intake shelters, full shelters (where a structured programme aimed at family reunification is offered), children's homes, alternative education programmes and prevention programmes. The goal of services was described as family reunification, with a children's home a last option if this failed.

Service providers nominated a number of factors as key to success. We integrate these into our recommendations section, for the sake of brevity.

3.9 Limitations of our work

Our work relied on focus group and individual interviews with children and young people in shelters and on the streets, and with the staff who served them, and a survey of over 300 street children in Gauteng Province. Both these methodologies have their limitations. Ethnographic and other more participatory methodologies could have given us a richer picture of street children's lives, and how they take to the streets and interact with services (Davies, 2008). However, we believe the study provides a valuable overview of the situation and could pave the way for further in-depth investigations.

4. Recommendations

Children living on the street may legally be considered ‘children in need of care’, but they are distinct in many ways from other such children. They have lived independently of adult care, and are often wary of adults and formal service provision (Davies, 2008). The most effective solution is to offer a range of services (Balachova et al., 2009) and to allow them to refer themselves, rather than forcibly placing them in care, as is done with other children found to be in need of care (such as children taken away from parents by the State because of abuse or neglect; National Association of Child Care Workers, 2005). Many of these services are already being provided by non-governmental organisations; however, attention from local and provincial government could ensure that the necessary range of services is provided, and that unproductive mistakes are not made. This will require cooperation across several departments: Social Development, Education, Health and others will all have roles. For instance, while Social Development might directly fund services, Education should be involved both in identifying children at risk (as we explain below) and in providing support to children returning to school.

4.1 Prevention

The presence of street children indicates that there is a larger population of children who are ill-nourished and under-educated because their families are too poor to support them adequately, or whose families are too violent to keep them safe (Donald & Swart-Kruger, 1994; Panter-Brick, 2002). Although preventive initiatives should seek to eliminate family poverty and family violence, these tasks perhaps lie beyond the scope of any one paper to describe. We therefore confine our remarks to identifying children specifically at risk of taking to the streets, even in a broader context of poverty and violence, noting that legislative support for preventive interventions is provided by the Children's Amendment Act (Act No. 41 of 2007).

One of the first tasks is to decide where to locate preventive initiatives, especially in a resource-poor environment such as South Africa. Clearly, since poverty is a major driver of children going onto the streets, interventions should be in poor areas, close to where children start their journey onto the streets, so as to intervene as early as possible. In Cape Town, service providers attributed a drop in numbers of street children in the city centre to this strategy. Appropriate areas may be identified via indices such as the South African Index of Multiple Deprivation (Noble et al., 2006), or by searching the records of those providing services to homeless children to identify the areas from which large numbers of children come. The latter procedure can be relatively easy if it forms part of a monitoring and evaluation plan in which all shelters participate (Ward, 2007).

Our respondents suggested that, once an area has been identified, specific children may be identified through their truancy from school. This naturally places a burden of responsibility on schools as the first line in a prevention system, as teachers should be the first to notice when a child is missing.

Assuming that programme resources exist, initiatives to support families should include at least the following, to be offered to families as needed: income generation and/or employment assistance, family violence interventions, parent training, substance abuse interventions, and support for children to receive an education. This last might entail assistance with school fees or uniforms but also, and equally importantly, assessing children for learning difficulties, and providing the necessary remedial support. These strategies directly address the reasons identified both by children in our study and by the literature (Swart-Kruger & Donald, 1994; Richter, 1996) for children taking to the street. Once a child is identified as at risk, a social worker or suitably trained interventionist should assess the needs of a particular family, and assist them, the school and their child to make use of the interventions they need.

Services that work with families and children at risk may prevent children from taking to the streets, and may keep them with their families of origin. As shelter managers observed, the longer a child has been on the street, the harder it is to intervene and the less likely it is that family reunification and educational interventions will succeed. This suggests that early rather than later intervention is more likely to succeed. At the time of our study, several shelters in Cape Town were initiating preventive interventions, but these had yet to be evaluated. This is clearly fruitful ground for future research.

4.2 Intervention

Although prevention is key, services should still be provided in the city centres as some children do come directly to the city on leaving home. However, city centres may be more likely to be the focus for politicised ‘quick fix’ initiatives, which service providers and the literature agree are unlikely to succeed, and may even aggravate the situation (Dybicz, 2005). Typical quick fixes include enticing children into programmes, removing children from the city centre to outer limits, or incarcerating them. Both the enticement and removal approaches are likely to erode children's trust in service providers, and neither is likely to achieve the long-term goal of reintegrating the children into mainstream society.

Specific recommendations included the need to provide services, along a continuum, from outreach work through soup kitchens and drop-in centres, to shelters and long-stay homes for children for whom reunification with their family is impossible. This enables children to develop trust in services that do not require a large personal commitment (such as a soup kitchen), and thereafter to develop the trust and courage required to make more serious commitments to shelters and long-stay programmes. This is a slow process that demands much of both the service provider and the child, but enables the child to attach to a programme and, through that attachment, to succeed in using the opportunities provided by the programme to achieve a more stable lifestyle. This is only possible if a suitable range of services, and suitable outreach workers, are available to the children in a city (Balachova et al., 2009).

Shelters should also collaborate. Staff members reported that, in their experience, children learn to play off one shelter against another if there is no collaboration. This tactic ultimately militates against the choices children need to make if they are to make the commitment to a programme. One example of such collaboration is that not all shelters in Cape Town have outreach workers: rather, outreach workers from one or two projects will make contact with children and refer them to the appropriate programme, thus reducing ‘competition’ between outreach workers for children.

Within the shelter, it is key that warm (yet boundaried) relationships are established with the children. Like any children, street children lack respect for adults they can manipulate. Relationships with children therefore go hand-in-hand with good collaboration between shelters, as well as between staff within the shelter. Warmth and limit-setting together enable children to develop trust in outreach workers and, through them, in the rehabilitation programme.

Also within the shelter, it is critical that shelters carry out thorough assessments once children enter a programme, to establish their physical and mental health needs and their educational needs, and to identify placement options for them. Following this, education appropriate to each child's needs must be provided (Donald & Swart-Kruger, 1994). Street children are highly likely to have experienced failure in the school setting, and to have one or more learning disability. Their educational needs must be assessed, and they should be placed in the appropriate setting. For some children, this will mean alternative educational programmes rather than mainstream schooling. Here the emphasis should be on functional literacy and vocational skills, so that they will gain the means to earn a living. Working with children to return them to mainstream schooling requires advocacy on behalf of the child with the school, as children may be old for the grade or have other special needs, and shelter staff said schools typically found it difficult to accommodate the special needs and sometimes difficult behaviour of street children. A significant gap they mentioned was children with mild mental handicap, who are not sufficiently handicapped for sheltered workshop employment but also cannot cope in mainstream schooling or employment.

Other needs that shelters struggled to meet were for psychiatric care and substance abuse treatment. Psychiatric illness (including substance abuse) in children can make their behaviour difficult to manage in the shelter, as well as form a barrier to placement. In addition, it was the impression of shelter staff that untreated mental illness among mothers had, in a number of cases, led to a lack of care for children, which had played a role in their taking to the street.

In terms of placement, shelter staff recognised that children who had a history of stable care, prior to taking to the streets were much easier to place with families. Their history of stable bonds with parental figures meant that they more easily formed attachments to guardians. Similarly, several street children had taken to the streets because of failed foster care placements, and these children may be more difficult to place – and for this reason shelters did not favour foster care. Nonetheless, both service providers and former street children said placements fail if families are not adequately assessed as being able to provide care for children. However, service providers said the majority of children were placed relatively easily with families, with two provisos: girls were easier to place than boys; and if the only family that could be identified was in a very rural area, the child's standard of living (relative to the shelter) might drop, and children might struggle with the placement for this reason.

Health and other services, as well as employment, are all more readily accessed when children have identity documents. Provision of these documents needs urgent facilitation, since their lack is a significant barrier to accessing services.

The five young women who had successfully completed shelter programmes nominated several additional aspects of shelters that helped them succeed. Shelters often begin to function much like families – several young women commented that they knew that, no matter what, they could return to the shelter if necessary. This was a significant support when they faced returning to their actual families and were unsure about how their families might receive them, or when they aged out of care (children in care are funded by a government grant until age 18). Other aspects of shelters that the successful graduates nominated as success factors were the schooling, educational outings and camps that the shelter made available to them. Volunteer programmes run by shelters allowed children to form relationships with adults other than shelter staff, which had been valuable to them emotionally.

There are numerous problems with children ageing out of care at 18 years old, as this typically means that they lose access to housing and education. This is likely to be similar for all children in care, and the situation urgently needs to be addressed as it may lead to children (whether they were originally street children or not) becoming homeless adults. This means that policy must be changed so that children can stay in care longer, and that funding must be provided once policy has changed.

4.3 Monitoring and evaluation

Monitoring and evaluation are key ingredients in evidence-based planning for services. Indicators for monitoring and evaluating the situation of street children and providing them with services have recently been developed (Ward, 2007). Setting up this monitoring and evaluation plan uniformly across a city makes it possible to assess the following areas of need:

• demand for beds in shelters (the quickest and easiest method for estimating the number of street children, without conducting a costly census);

• mapping where children come from, in order to target preventive services appropriately;

• ensuring quality services by registering and monitoring service providers;

• ensuring service access by making sure that services are located where they are most needed; and

• ensuring service access and quality by monitoring the health and education needs of children, whether these are met, and why they are not met.

While a census or an in-depth study can provide a great deal of information in a single snapshot, monitoring and evaluation can provide more information, over time, at a relatively low cost, and make it possible to assess trends in services and in the population of street children. It can also assess whether services are functioning, and whether children are getting the help they need. Such a system is essential to evidence-based planning.

5. Summary

The number of children on our streets is, in essence, an indicator of the number of children living in poverty. The higher the numbers of street children, the more families are struggling to support and provide for their children. While intervention services may always be needed, the number of children on our streets can only be reduced through preventive initiatives – by reducing the numbers of families who live in poverty, supporting families so they can raise their children in safe, loving environments, and helping children stay in school. Clearly, attention needs to be paid to this at the broader, systems level – through initiatives that seek to reduce poverty in general. In the meantime, however, initiatives can be targeted towards children most at risk: those living in poor areas who begin to play truant from school.

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 Human Sciences Research Council, the Gauteng Department of Social Development and the Swiss Agency for Development and Cooperation is gratefully acknowledged.

The authors wish to thank the young people, former street children, and service providers who took time to tell us about their lives and work, without whose generosity the study would not have been possible, Human Sciences Research Council staff members Annemarie Booyens, Anneke Jordaan and Monika Peret, and our fieldworkers for hard work under trying conditions, Tsiliso Tamasane for help with interviews, and Michael O'Donovan for finding a way to estimate the number of street children in Gauteng.