CYC-Net on Facebook CYC-Net on Twitter Search CYC-Net

Join Our Mailing List

52 MAY 2003
ListenListen to this

child labour

A broader view of poverty

Rachel Bray, Research Fellow in the Centre for Social Science Research, University of Cape Town, looks at what social surveys tell us about child work in South Africa.

This article draws attention to one aspect of child well-being that is poorly understood in South Africa, namely children's working roles and the consequences of work on their own well-being and that of the household. The discussion is drawn from a longer paper that examines our knowledge of a wide range of factors influencing children's economic, physical and psycho-social well-being (Bray 2002). Its aim was to assess the available survey data on children's lives in South Africa in order to see whether we have the necessary tools to trace changes in child poverty and well-being over time, and to link these changes to broader social, political and economic trends.

Child poverty and economic well-being
Most reports of children's economic well-being focus exclusively on child poverty as estimated through surveys of household poverty. For example, recent estimates of child poverty trends using the OHS1 household income data and poverty line of R400 per month per child found that 75.8% of South African children lived in poverty in 1999 as compared to 64.7% in 1995 (Streak, 2002:3-4). The general picture painted by such calculations is important, however they do not tell the whole story of children's economic environments, for three reasons.

Firstly, calculations based on household income alone do not give an accurate picture of child poverty and its impact on well-being. A better picture is formed by using a composite index of poverty indicators and an appropriate data set. For example, Haarmann (1999:28) used the 1994 PSLSD2 data to calculate that 69.3% of children aged 0 to 6 years lived in poverty, meaning a lack of economic means and access to basic services (such as adequate shelter and education).

The second reason is that they do not take into account dynamics within the household that influence the allocation of resources. Therefore, these calculations of child poverty reflect the household resources that are theoretically available to children, rather than actual provision for children. Currently, we lack information on the resources allocated to children and other members of the household.

The third factor not accounted for in calculations of child poverty is children's contributions to the household economy and/ or their own livelihoods.

Moreover, the costs or benefits to child well-being that result from different forms of work are poorly understood. The remainder of this article explores these two aspects of child well-being.

Understanding the economic roles of children
Owing to the official definition of the working age population as those between 15 and 65 years, most national household surveys do not collect data on employment on anyone under 15 years. The few exceptions are the 1995 OHS, which asks for employment information on those aged 10 years and over, and the PSLSD, which captures information on adults' and children's involvement in casual/temporary work and self-employment.

Often, the reasons why existing national surveys cannot yield accurate information on child work and its effects on well-being are related to the way in which questions are worded and asked. For example, the Labour Force Survey question on school attendance is posed in such a way that it is likely to elicit information on who is enrolled and therefore should be attending, rather than the actual attendance rates of children or their achievements3. Hence the information is inadequate for analysing the effect of long working hours on children's education. Another example is the question in the IES4 on the income of every household member, which should tell us about the income generated by children (through earnings, or work in kind for family gain). However, accurate information is unlikely, because children are very unlikely to be responding for themselves, and adult respondents may be unaware of their income, or consider it insignificant,5 or not wish to disclose income from children under 15 years old because work is illegal for this age group.

In searching for information on children's work, I am struck by the diversity in estimates of the numbers of working children in South Africa. Bosch and Gordon's (1996) analysis of the October Household Survey of 1994 finds that approximately 200 000 children between 10 and 14 years (or 4% of that age group) are working. However, the PSLSD and OHS 1995 identified very small numbers of employed children under 14 years (ten and nine individuals respectively) (Muller, 2002:11-12), meaning that these figures cannot be extrapolated to provide a national picture. There are three plausible explanations for the discrepancy between figures relating to child work. The first is the use of different definitions of work. In other words, if work is considered to be any form of productive contribution (remunerated for cash or kind, or non-remunerated) to the family livelihood including domestic work, then figures for “child workers” will be much higher than if a narrower understanding of employment for remuneration is used. The second explanation is that the wording of questions and the context in which they are asked inhibit responses relating to children's work. The third explanation is the different reference periods used for measuring children's participation in work.

Looking firstly at the definition of work used in national surveys, we find that although there is no formal exclusion of children, they are excluded owing to criteria relating to work. For example, the Mesebetsi Labour Force Survey (1998) collected data on the employment, demographics and wages of all household members. It included a question asking whether each person spent time collecting wood or water, occupations we know are often performed by children. However, these data are not included in measures of employment because “although these activities do contribute to the household economy, they were not classified as employment unless done for other households for pay”. (Fafo, 2001:6).

Given that South African children are unlikely to be working for an income over a sustained period, the length of the reference periods used will strongly influence rates of child work captured. Most household surveys classify people as employed if they have worked in the seven days prior to the interview (Muller, 2002:11). The PSLSD questionnaire asks for information on employment of all household members over the past month. However, the more recent Survey of Activities of Young People (SAYP) uses a reference period of twelve months, specifically in order to capture seasonal or irregular child work (Orkin, 2000). The outcome is that any child who had conducted one of a range of activities designated as work (including domestic and agricultural tasks with or without income) for even a few days could be classified as a working child. As we would expect, the results suggest that much higher numbers of children are working than indicated in any preceding survey.

It is important to remember that the SAYP was a collaborative endeavour of the Department of Labour and ILO-IPEC6 designed to fill gaps in data on the nature and prevalence of child labour in South Africa. In addition to internal demands for more information on child work7, the survey was influenced by the priority of ILO-IPEC and international donor agencies, namely to gather information that can assist in the identification, and eventual elimination, of hazardous child work. Owing to the fact that this definition of child labour does not limit “work” to work for financial or other gain, but includes activities such as domestic tasks that could potentially have negative effects on their development, the SAYP set out to gather detailed information on a wide range of children's activities8. The results of the SAYP most relevant to our discussion are as follows:

26% of South African children had been economically active (defined as economic activities for pay, profit or family gain, and excluding unpaid domestic work and fetching wood and water, for any amount of time per week),

45% of South African children are engaged in child labour (defined as a minimum of 1 hour of economic activity per week, and/or 5 hours of school labour and/or 7 hours of household chores),

36% of South African children are in the higher risk category of child labourers (preferred by the Department of Labour), because they spend at least 3 hours per week engaged in economic activities, and/or 5 hours in school labour9 and/or 7 hours doing household chores (Orkin, 2000).

Before we can use these findings to draw conclusions about the impact of work on child well-being, we must examine them closely with respect to the nature and frequency of children's work. For example, the definition of economic activity used includes unpaid domestic work for nonfamily members and fetching wood/water, but excludes other chores within children's own households. If we look at a breakdown of the most common types of work done by South African children, we find that fetching wood/water is the most common, followed by farm work, followed by very low levels of paid work in other sectors. It would therefore appear that for the vast majority of children, what we are actually talking about is the impact of work in and around the home and farm on their wellbeing. Only a very small proportion of this is likely to be immediately hazardous to child health (for example exposure to toxic pesticides on commercial farms), and the pertinent questions lie in the extent to which children are able to balance a working role and educational progress, and whether children are at greater risk as a result of their working role (through gender, age, birth order or unusual family composition). The proportion of time children spend working is of course relevant to impact on well-being10. Few of the jobs done by the children surveyed are likely to have significant impact on children's well-being if done for less than three hours per week.

In conclusion, national household surveys, by their nature, are primarily interested in the socio-economic status of households and, owing to the methodological limitations outlined above, effectively exclude children's contributions to their own well-being or that of the household. In contrast, the SAYP focuses specifically on children and sets out to identify children's activities that can be classified as work, and therefore can draw exaggerated conclusions around the nature and effects of children's work. Neither approach is able to capture the way children experience and perceive work, nor the economic and social gains felt by parents for whom the work contributions of children make a critical difference to household functioning.


Bosch, D and A. Gordon. (1996). Child labour in commercial agriculture in South Africa. Paper prepared for the ILO Regional African Workshop on Child Labour in Commercial Agriculture, Dar es Salaam, Tanzania 27-30 August and for a study of the South African Labour Market by the ILO Labour Market Policies Branch.

Bray, R. (2002). Missing Links? An examination of contributions made by social surveys to our understanding of child well-being in South Africa. CSSR Working Paper No. 23. Social Surveys Unit, Centre for Social Science Research, University of Cape Town.
Fafo. 2001. Introduction to the Mesebetsi Labour Force Survey.

Haarman, D. (1999). The Living Conditions of South Africa’s Children. Research Monograph No. 9. Applied Fiscal Research Centre (AFReC), University of Cape Town.

Muller, C. (2002). Measuring South Africa’s Informal Sector: An Analysis of National Household Surveys. Paper presented at the Development Policy Research Unit (DPRU) Second Annual Conference on Labour Markets and Poverty In South Africa, Johannesburg 22-24 October 2002.

NPA. (2001). Children in 2001: A Report on the state of the Nation's Children. National Programme of Action for Children in South Africa, The Presidency.

Orkin, F. (2000) Child Labour in South Africa: Tables. Survey of Activities of Young People 1999, Statistics South Africa, commissioned by the Department of Labour, South Africa.

Statistics South Africa (2002). General Household Survey: Modules.

Streak, J. (2002). Provincial child poverty rates and numbers based on poverty lines of R400 and October Household Survey Data. Child Poverty Monitor No.1. August 2002. IDASA.


1. October Household Survey
2. Project for Statistics on Living Standards and Development
3. Problems arising from the phrasing of questions relating to education are discussed further in part 3.
4. Income and Expenditure Survey (1995 and 2000)
5. The questionnaire is administered through face to face interviews with the household head or responsible adult (Statistics South Africa 2002:4).
6. The International Labour Organisation's International Programme for the Elimination of Child Labour
7. The lack of adequate and reliable data on the nature and extent of child labour prior to the SAYP is stated in the government’s 2001 Report on the State of the Nation's Children (NPA, 2001:113).
8. In order to identify the proportion of this activity to be classified as “child labour”, a number of proxy indicators and filters were used to qualify the findings.
9. School labour is defined as school maintenance and improvement activities, for example cleaning toilets (Orkin, 2000).
10.It is worth bearing in mind that the figure of 26% economically active children is reduced to 15% if the criterion of at least one hour of work per week is applied, and to 8% if the cut-off is set at least 3 hours per week.

This feature: Bray, R. (2003) A broader view of poverty. ChildrenFIRST, April/May 2003

The International Child and Youth Care Network

Registered Public Benefit Organisation in the Republic of South Africa (PBO 930015296)
Incorporated as a Not-for-Profit in Canada: Corporation Number 1284643-8

P.O. Box 23199, Claremont 7735, Cape Town, South Africa | P.O. Box 21464, MacDonald Drive, St. John's, NL A1A 5G6, Canada

Board of Governors | Constitution | Funding | Site Content and Usage | Advertising | Privacy Policy | Contact us

iOS App Android App