Child-headed households in South Africa: What we know and what we don't

Abstract

Household structures in many countries in Africa have been drastically changed by the HIV/AIDS epidemic. Traditionally, orphans were absorbed by members of the extended family but today this safety net no longer works so well and many children are left to fend for themselves. The South African Government recognises the problem, but views differ on the effectiveness of the programmes for assisting these children. This article looks at what is known and what needs further research. It is known that child-headed households do exist in South Africa and that children living in these households are vulnerable. What is not known is the prevalence of these households and how to deal with the children's psycho-social problems. The article argues that the definition of child-headed households needs to be refined if we are to understand the extent of the problem. This will facilitate assessment of the programmes established for these children.

Keywords:

• HIV/AIDS
• child-headed households
• orphans

1. Introduction

As the HIV/AIDS epidemic matures, a variety of negative social consequences have been observed. Sub-Saharan African countries affected by the epidemic are experiencing a significant change in the types and composition of households (see for example Heuveline, 2004). Of particular concern are the child-headed households, defined as those that are under the care and supervision of a person under 18 years old. Such households are not something new in Africa, but traditionally there were safety nets that prevented them from being formed or dissolved them as soon as they were created. In some African settings it was the custom for relatives, immediately after the funeral, to openly discuss the fate of young children who had lost their parents. The elders would make sure that all minors were placed in households where there was an adult capable of taking care of them. But today, with the high frequency of adult deaths, the traditional safety nets can no longer cope. The extended family system can no longer absorb all the children who have lost the care and support of parents. Child-headed households are therefore becoming a problem, and how to deal with the problem is a matter for debate.

In South Africa, views of the problem were at first divided. One group, mainly non-governmental organisations (NGOs), faith-based organisations (FBOs), community-based organisations (CBOs), civil society organisations (CSOs) and development partners (such as the UN, the World Bank and the IMF), recognised that child-headed households do exist and that this is a problem to society. The other group, academics and researchers, were cautious in discussing the matter as they believed that the problem of child-headed households was not as extensive as was claimed (Desmond et al., 2003). Currently it is being argued that the problem has been given more emphasis than it deserves. Some researchers argue that the number of these households is small and they are not the poorest households (Richter & Desmond, 2008). Given the South African Government's current emphasis on child-headed households, there is a need to investigate the facts before cutting down on the existing efforts.

This article examines the situation via a review of the literature. It looks at what is known, points out areas where more research is needed, and suggests a way forward. The specific objectives are:

•	to sharpen the definition of child-headed households,
•	to discuss the nature and consequences of child-headed households,
•	to comment on recently published estimates of the prevalence of child-headed households, and
•	to suggest future possible avenues for researchers and policymakers.

The review of the literature covers publications and research projects, some of which used secondary and others primary data. The article also draws on the findings of a research project that investigated the changing family structures in South Africa by collecting both quantitative and qualitative data from three provinces, Limpopo, KwaZulu-Natal and the Eastern Cape (Mturi et al., 2005a, 2005b). Although the project was not designed to deal only with child-headed households, it was key in providing a crucial understanding of the subject matter and revealed how serious the problem of child-headed households is.

2. Definition of a child-headed household

Definitions of ‘child-headed household’ vary, and depend to some extent on the definition of ‘child’. Some authors define a household as ‘child-headed’ if the head is an orphan or attending school. The national study by Mogotlane et al. (2010), commissioned by the Department of Social Development (DSD), included adults as old as 34 and well over half of the sample were 18 or older (Mogotlane et al., 2010).

The age at which a person is considered to be a child is clearly stated in the South African Constitution: it is a person who has not yet turned 18 (RSA, 1996). Therefore, categorising people older than 18 as ‘children’ is incorrect in the South African context. The lowest age used has also been a problem, as the number of child-headed households becomes quite small at very young ages. The 2001 census, for example, suggests the lower bound should be restricted to 10 year olds (Stats SA, 2004), since there are not many households headed by children under the age of 10. This lower age boundary creates a bias and could underestimate the prevalence of these households. For instance, Ruiz-Casares (2009) reports that children interviewed as head of household in her study conducted in Namibia were as young as nine. It is argued here that the age-based definition of a child-headed household should consider the age range nine to 17.

A further complication is what has recently been referred to as the ‘child only’ household (Richter & Desmond, 2008; Meintjes et al., 2010). These are households composed only of children. Obviously this constitutes the bulk of child-headed households, but this narrows the definition as it excludes households headed by children that contain adults who are either very old or too sick to be responsible for the household (Van Dijk, 2008; Ruiz-Casares, 2009).

This article supports the definition suggested by the DSD that says a child-headed household:

is recognised when the parent or primary caregiver of the household is terminally ill or has died, no adult family member is available to provide care and support and where a child has assumed the role of primary caregiver in respect of a child or children in the household in terms of providing food, clothing, and psycho-social support. (2005:5)

This definition accepts the fact that a child-headed household may have adult members but the key issue is that a child is the one responsible for the day-to-day upkeep of the household. Statistics South Africa defines a household as a person, or a group of persons, who occupy a common dwelling (or part of it) for at least four days a week and who provide themselves jointly with food and other essentials for living (Stats SA, 1998:12). This article proposes therefore that studies of child-headed households should use the following definition:

A child-headed household is one in which a person or a group of persons (regardless of their age) live together for at least four nights a week, eat together and share resources, and a person under 18 years of age is responsible for the day-to-day upkeep of the household.

During data collection it is desirable that the head, whether it is a child-headed household or not, identify himself or herself after the interviewer or enumerator has provided the necessary definition of ‘head’. In the absence of the head, other members of the household can identify this person.

3. The existence of child-headed households in South Africa

There is consensus that child-headed households do exist in South Africa. Evidence from the major nationally representative surveys has been presented by Richter & Desmond (2008) and Meintjes et al. (2010). Although it is argued that their estimates are low (details are presented in Section 5 below), this does not alter the fact that these surveys confirm the existence of these households in South Africa. A more recent situational analysis of child-headed households was conducted in all provinces of South Africa in 2008 (Mogotlane et al., 2010). Although this study uses a different definition of ‘child-headed household’ (as explained above), it leaves no doubt that the phenomenon exists in all nine provinces. Other recent field studies on the topic are HSRC (2002), Guest (2003), Shisana et al. (2005), Madhavan & Schatz (2007), Maqoko & Dreyer (2007) and Van Dijk & Van Driel (2009).

In 2003 a research team conducted a qualitative study in Limpopo, KwaZulu-Natal and the Eastern Cape of the types of households in South Africa, the pressures they face in the light of HIV/AIDS epidemic, the changes that have occurred over time and the coping mechanisms they use (Mturi et al., 2005a). It was clear from the first phase of the study that there were families comprised only of children, where an older child was looking after the younger siblings. Respondents reported that child-headed households existed in their areas, and researchers observed such households personally, especially in the Eastern Cape. Meintjes et al. (2010) state that about 90% of the child-headed households in South Africa are located in these three provinces.

The following are some examples recorded by Mturi et al. (2005a) of what respondents had to say about the presence of child-headed households. One female respondent in the rural Eastern Cape said:

Our neighbours at the back are children who lost their mother a while ago and were left with their father, but last week we buried the father, so they are left on their own.

An elderly female respondent in rural KwaZulu-Natal was concerned that there are children living on their own and not getting enough assistance:

There are my brother's children whose father passed away and their mother followed him immediately. They are living on their own. We go to check them but we do not have the ability to help them. We also do not have people who help us.

One respondent was a male child found living on his own in the rural Eastern Cape. When invited to talk about his living arrangement, and whether that was what he wanted, he said:

I stay alone – my parents passed away. I have an uncle in Quzini but he treats me badly, together with his wife.

The second phase of this study was a quantitative survey conducted in 2004 in the same three provinces. It involved a non-probability sample of over 1000 households and data were collected from all types of households (Mturi et al., 2005b). Seventeen child-headed households were identified in this study, headed by children as young as 11.

4. The causes of child-headed households and the needs of children living in them

The main cause of child-headed households is the death or illness of a parent or parents. Sometimes the bereaved children are accommodated by members of the extended family, such as aunts or uncles or grandparents, but then form a child-headed household when these guardians become ill or die. Besides the effect of the HIV/AIDS pandemic, another cause that has been identified is labour migration, especially in South Africa. Parents may migrate for work-related reasons and leave their children behind. During the apartheid era, migrant African workers could not take their families with them to the cities (Posel et al., 2004). Hence, there was a tendency to leave behind children under the care of an older sibling or a relative, since mothers also migrated to the cities, often to work as domestic workers. Child-headed households caused by migration of parents are usually better-off as there are remittances sent which give the children an economic base. Parents also attend to serious problems like illnesses and request other relatives or friends to keep an eye on the children. The child-headed households that existed during the apartheid era (prior to 1994) because of the migrant labour system were thus not known to be a serious problem to society. Some child-headed households are still of this kind today, and similarly are not a serious social problem. Studies need to distinguish between these and the kind that result from loss rather than just temporary absence of parents. Some child-headed households may of course be caused by the disappearance of a parent or parents, but this is of course uncommon.

The particular interest in studying child-headed households is that here we find children doing what children are not expected to do. They are rather expected to be in school and receiving the care and support of adults. There is a notion that children living in impoverished households are all the same regardless of who is heading those households (Richter & Desmond, 2008). This article argues otherwise. Even if they are in the same situation economically, children's needs are for more than just money. There are things which are equally important, such as love and belonging, safety and security (DSD, 2010). Donald & Clacherty (2005), in their KwaZulu-Natal study, compared children living in child-headed households with children living in adult-headed households in equivalent impoverished communities. They found that those in child-headed households were more vulnerable than their counterparts in adult-headed households to such problems as failure to access social services, inability to generate resources, unresolved grief, lack of attainable long-term goals, poor self-worth and poor internal locus of control. There is a need for further research to help understand the differences between children living in child-headed households and their counterparts living in poverty in adult-headed households. This is very important for policy as well as for identifying resources that will assist these children.

Since most child-headed households are formed due to the death of a parents and one of the main causes of early adult death is HIV/AIDS, most of these children are also AIDS orphans. These children often live through the experience of dealing with parental illness, and for many the experience of carrying the responsibility for a household begins some time prior to being orphaned. Because of the stigma and discrimination that so often surrounds HIV/AIDS, the children are often not aware of a parent's HIV-positive status. Parental death from AIDS, resulting in the children being orphaned, ‘marks a major crisis in the already existing trauma of a parent's illness and death, emotional devastation, economic decline and often social ostracism’ (Foster & Germann, 2002). An earlier literature-based study (Sengendo & Nambi, 1997) shows that children orphaned and rendered vulnerable by HIV/AIDS exhibit high levels of depression, emotional distress and conduct problems, and that these psycho-social adjustment difficulties appear to be more severe than for children orphaned by other causes or children whose parents are still alive but absent.

Parental bereavement from HIV/AIDS has a number of socio-economic and psychological consequences. It has been widely documented that children orphaned by HIV/AIDS have reduced opportunities for schooling and suffer physical hardships such as shortage of food and lack of social security (Nyambedha et al., 2003; UNICEF & UNAIDS, 2004). Reduced access to food manifests in various forms, such as lack of proper and nutritious diet and insufficient daily meals. Nkomo points out that:

Anecdotal South African media publications report that children orphaned and rendered vulnerable by HIV/AIDS also have poor access to social grants, due to a number of factors, such as lack of birth registration documentation, delays experienced in the processing of grants at the Department of Social Services, and so on. (2006:30)

Nkomo goes on to say that ‘children orphaned by the HIV/AIDS pandemic also suffer from a number of psycho-social consequences’, including ‘grief, loss of identity, such as the self, family and cultural identity, dealing with shame, stigmatisation and abandonment as well as rejection and death’ (2006:30).

All these problems that affect orphans are even more serious for the orphans who live in child-headed households. Maqoko & Dreyer list the challenges that child-headed households face:

a serious threat to education because of poverty, difficulties in obtaining food and shelter; a high risk of being sexually abused by relatives and neighbours; the threat of child prostitution and child labour; difficulties in getting birth registration which is a prerequisite in procuring healthcare and social security benefits; and experiencing property grabbing by families and communities. (2007:724)

5. Levels and trends of child-headed households in South Africa

It is important to understand the prevalence of child-headed households for many reasons. The government is providing considerable funds to address this problem and yet the magnitude of the problem is not known. This will be money wasted if the problem is not as serious as it is understood to be by the government. Knowing the extent of the problem will either justify the use of resources by the government if the prevalence is high or prompt a rethink if it is low. Some researchers argue that the problem of child-headed households has been over-emphasised by the government (Richter & Desmond, 2008). The only way to know which side is correct is to measure the prevalence of these households accurately.

Table 1 shows various estimates of the percentage of child-headed households in South Africa, based on the nationally representative surveys and censuses conducted by Statistics South Africa. It is clear that there are serious disparities between these estimates, which vary from as low as 0.11% in 1995 to as high as 0.89% in 2001. The 2001 census indicates that of 11.2 million households enumerated, nearly 100 000 were headed by children. This is a significant figure for society to worry about. As explained in Section 7 below, the main reason for the differences is that the data from the household surveys were not designed for providing estimates of child-headed households. Another reason is that the definitions of child-headed households used were usually inadequate (Van Dijk, 2008). Given the variations observed in the estimates, we can reasonably conclude that the exact percentage of child-headed households in South Africa is unknown.

Table 1: Estimates of the percentage of households in South Africa headed by children under the age of 18, as published by various sources

Source	Percentage of child-headed households	Type of Stats SA data	Reference year
Richter & Desmond (2008)	0.11	OHS	1995
Richter & Desmond (2008)	0.34	OHS	1997
Richter & Desmond (2008)	0.45	OHS	1999
Computed from Stats SA (2004) tables	0.89	Census	2001
Richter & Desmond (2008)	0.67	GHS	2002
Richter & Desmond (2008)	0.66	GHS	2003
Meintjes et al. (2010)	0.47	GHS	2006
Note: OHS = October Household Survey, GHS = General Household Survey.

6. Recent policy and programme development to assist child-headed households in South Africa

The problem of child-headed households was first identified in sub-Saharan Africa in the late 1980s (Foster & Makufa, 1997). In South Africa during the 1990s it was mainly the concern of activists who were trying to sensitise the government and the public at large to the problem. At the centre of activism were NGOs, FBOs, CBOs and CSOs. The first three were, and still are, central in providing necessary services for children living in child-headed households. It took some time for the government to respond, but tremendous effort has been directed to these households since 2000. A major step was taken by having a regional SADC conference in Pretoria, 2 to 5 June 2002, organised by the South African Department of Social Development. The conference was entitled ‘A Call for Coordinated Action for Children Affected by HIV/AIDS’. Its theme was ‘Strengthening coordinated action for orphans and other children made vulnerable by HIV and AIDS’. Children living in child-headed households are of course part and parcel of the category ‘orphans and vulnerable children’. One of the main recommendations of the conference was for the Government of South Africa to establish a coordinating structure for these children at national, provincial and district level. Currently, there is a permanent national coordinating structure for orphans and vulnerable children, the National Action Committee for Children Affected by HIV and AIDS (DSD, 2005). All the provinces and their districts were supposed to establish a similar structure that would coordinate at provincial and district levels, but the process of implementation has been very slow.

Another major achievement by the Government of South Africa on matters to do with child-headed households is the Children's Act (RSA, 2005). This Act is one attempt to look after the interests of all children, in line with the ‘Guidelines for the Alternative Care of Children’ adopted recently by the UN General Assembly (UN, 2010). The South African Children's Act has three special provisions that were instituted for children living in child-headed households. The first has to do with recognising that these households do exist. The second states the intention of having an adult supervisor designated to look after children in these households. The third has to do with care and protection for children living in these households whereby a designated social worker is given instructions on how to deal with these children (RSA, 2005). Although there are doubts about the implementation and monitoring, especially when it comes to allocating supervisors (see for example Van Dijk & Van Driel, 2009), this is a major step for a national government and the provisions of the Act need to be carried out.

The Children's Act was followed in 2008 by a situational analysis of child-headed households in South Africa, a national study conducted under the auspices of the DSD (Mogotlane et al., 2010). This was one of the government's commitments to investigating these households and the circumstances in which the children involved live. The findings of this study have been presented in various provinces in order to make government officials and other stakeholders aware of the challenges of child-headed households. Recommendations arising from this analysis were that there is a need to:

•	develop a national strategy and action plan on child-headed households;
•	conduct workshops and training on legislation and policies relevant to child-headed households;
•	rigorously monitor and evaluate social grants to ensure that children who qualify access the grants;
•	set up skills development programmes for child-headed households and provide counselling to build courage, self-esteem and confidence; and
•	strengthen and formalise collaboration between departments and sectors.

The government effort to assist child-headed households with accessing social grants is important. Two grants are available to these children: the child support grant and the foster child grant. Work still needs to be done to make these grants effective in assisting these children and the government is revising the conditions for acquiring these grants. Recently, the government has produced the National Guidelines for Statutory Services to Child-Headed Households (DSD, 2010). This document gives more details on the nature and the needs of child-headed households. These efforts by the government need to be complemented by accurate and relevant research that will provide a better understanding of children living in these households. The information given above suggests that the development of policies and programmes on child-headed households is quite advanced in South Africa, but there is a need also to understand what is happening on the ground. How the policies and programmes actually translate into improving the life of these children is beyond the scope of this article.

7. Use of household surveys to estimate prevalence of child-headed households

Nationally representative household surveys, such as the October Household Survey and General Household Surveys, conducted by Statistics South Africa are complex sample surveys that gather information relevant for providing demographic and socio-economic indicators for the country. By nature, they are large, complex and expensive. The sampling design is done by experts who make sure that the sample is randomly selected so as to represent the whole country. Although the primary sampling unit is a household, the sampling procedure used does not deal with different varieties of households. In other words, the type of household that is more common will tend to overshadow the less common types. These general purpose surveys may thus seriously underestimate the prevalence of child-headed households. This view is supported by Monasch & Boerma (2004), who note that households without an adult member may not be enumerated in a household survey because generally an adult is required to complete the household questionnaire and this will cause under-representation of child-headed households. Van Dijk note that ‘[t]he main problem with estimates of the number of child-headed households in South Africa is that they are based on household surveys which are not constructed with the objective of determining the number of child-headed households’ (2008:6).

It can be argued that recent estimations of the levels of and trend in child-headed households using the recent nationally representative surveys (see Table 1) have improved our understanding of the nature and extent of child-headed households in South Africa. But what the statistics tell us may not necessarily be true, as an example will show. In demography, various methods have been used to measure the level of maternal mortality. One popular source of data is household surveys. But as the World Health Organization states:

maternal deaths are rare events in epidemiological terms, surveys to measure their levels require large sample sizes to provide statistically reliable estimates … Even with large sample sizes, the obtained estimates are still subject to uncertainty (wide confidence intervals) making it difficult to monitor changes over time.

(WHO, 2007:6)

Indeed, the over-sampling that has been done over the years by the countries participating in the worldwide Demographic and Health Surveys programme has not produced estimates of maternal mortality that are very reliable.

Child-headed households are also rare in the sense that they cannot be compared to other types of households such as households consisting of a nuclear family. There is a need to conduct a household survey that takes into account the actual composition of the households at the sampling stage. In other words, the sampling design must take into account the nature, extent and rarity of the child-headed households when designing a sample that is meant to provide reliable estimates of these households. There is still no certainty about the number of child-headed households in South Africa. Researchers continue to try to understand what has been achieved and what is still to be achieved to help child-headed households, but their efforts are hampered by not having reliable figures on the prevalence of these households.

8. Discussion and conclusion

All stakeholders agree that child-headed households exist in South Africa. These are households whose members are all children under 18, one of whom (usually the eldest) becomes responsible for the rest, or households whose adult members are unable to head the household, so a member under 18 takes that position. Obviously these households have many problems arising from limited income (such as shortage of food, inadequate clothing, lack of health care), reduced opportunities for education, housing problems if the parents did not have shelter, and psycho-social problems. There is a need for a systematic way of assisting children in these households. The South African Government has gone a step in the right direction in acknowledging that child-headed households do exist and that children living in these households are in difficult circumstances.

When parents die and children are left alone, a limited number of options are available for these children. The following are three possible options. The first, and perhaps the best, is the traditional African practice where the extended family assumes responsibility. Unfortunately, as mentioned, the increased number of these households means this solution no longer functions as well as it did. The second is to put these children together in institutions such as foster care schemes and orphanages. This approach has many negative consequences and has been widely criticised (Richter, 2008). The third, the one adopted by the South African government, is to let these children live in surroundings that they are familiar with and be assisted there. This approach has the advantage of letting them live together as a family.

The Children's Act makes special provisions for children living in child-headed households. This initiative needs to be welcomed as it was achieved as the result of a long battle by activists comprising NGOs, CBOs, FBOs, CSOs and the media. Instead of claiming that the government is over-emphasising the problem of child-headed households, we need to welcome this achievement and assist it further so that these children can be empowered to lead normal lives. Two matters from the Children's Act in relation to child-headed households need to be analysed further. The first is the idea of child-headed households functioning under the general supervision of an adult designated by a children's court or by an NGO. With all its good intentions, this provision could be difficult to put into practice as it assumes that all adults given this task will be responsible. Very clear instructions will be needed on how to implement this initiative. The second is the important matter of helping children in child-headed households to access the grants to which they are entitled, which will at least give them financial assistance.

The recent literature has presented estimates of the prevalence of child-headed households in South Africa using the nationally representative surveys that targeted all types of households. As an initial step, these estimates are helpful for stimulating discussion on the extent of the problem, but they should not be taken as final, since the surveys they are based on were not designed specifically to count child-headed households. There is urgent need for a survey designed for this purpose, using a good definition of a child-headed household. The survey should distinguish between child-headed households resulting from the death or sickness of parents and those resulting from the parents working elsewhere. It should also provide indicators for monitoring and evaluation purposes that could be used to assess the effectiveness of the programmes established for these children. Such a survey would cater for various research purposes. First and foremost, it would provide a more accurate estimate of the prevalence of child-headed households at national and provincial levels. Secondly it would provide much-needed information about the circumstances of children living in child-headed households and the difficulties they face. All initiatives (by the government and otherwise) to assist these children should continue as planned while waiting for the results of such a survey.