Richard ANKER [*]
Ill-fed and ill-clothed working children from developing countries are frequently depicted on television and in the print media. At the beginning of the new millennium, the work done by these unfortunate children is an unacceptable aspect of life in all too many countries.
This condemnation of child labour by society coexists with other seemingly contradictory attitudes. First is the fact that many children work willingly and with their parents' support. If child labour is so bad for children, why then do so many parents allow or encourage it and why do so many children willingly engage in it? In developing countries, the usual explanation for this apparently irrational behaviour is poor families' need for additional income to help ensure their survival. Yet, though this reasoning has considerable merit, it does not explain why the incidence of child labour varies across poor households within communities, across poor communities within countries, and across poor countries throughout the world. Second, in the right circumstances it can be good for children to work. For example, there is widespread agreement that non-hazardous forms of work can teach children self-reliance and responsibility. Indeed, it is common for children in high-income countries to work, usually to earn th eir own pocket money -- e.g. doing babysitting, delivering newspapers, or working in their family's business or farm; as well as in restaurants or shops after school and during school holidays.
This article develops a conceptual framework within which to situate the economics of child labour, the aim being to address seemingly contradictory aspects of the phenomenon, such as those noted above. Based on this conceptual framework, implications are drawn for the measurement of child labour, as well as for policies and programmes for addressing child labour and children's welfare. Thus, the following questions are addressed in this article:
* What is child labour? Why should one be concerned about it?
* How is child labour presently defined and measured? How should it be defined and measured?
* What are the policy implications of an economic analysis of child labour based on the conceptual framework developed here?
The underlying contention is that child labour policies, programmes and research have tended to be excessively simplistic, paying too little attention to the complexities of the phenomenon. This has retarded the identification of appropriate and effective policies and approaches for the elimination of unacceptable forms of child labour. For these reasons, the conceptual framework developed here takes into consideration that the economic benefits and costs associated with the elimination of child labour are influenced by three factors: (i) child labour exists in various forms; (ii) there are several ways to justify the elimination of child labour; and (iii) a wide range of actors and institutions would be affected by the elimination of child labour.
The article starts with an examination of the reasons for being concerned about child labour. The next section considers measurement issues and the need for a number of different estimates of child labour in order to distinguish between its various forms.  The final section draws out policy and programme implications for those interested in the elimination of child labour.
Reasons to be concerned about child labour
Before considering issues of definition and measurement, it is best to start at the conceptual level, by identifying the aspects of child labour which are of overarching concern:
* the protection of children;
* the development of children;
* economic and labour market impacts of child labour.
Table 1 lists the main reasons for concern, possible indicators with which to measure them, and selected comments.
Protection of children
Protection of children is the primary reason why many people and organizations are concerned about child labour. Children are vulnerable in a number of ways, and childhood is a period of life during which special protection is needed; furthermore, children who work are often exposed to abuse and exploitation. An international consensus has now emerged against what are regarded as especially unacceptable forms of child labour. The ILO's Worst Forms of Child Labour Convention, 1999 (No. 182), reflects this concern and the international consensus committed to the elimination of child labour.
This concern for children's welfare is mainly humanitarian in nature, although there are also associated economic concerns, for example, the impact on health and other costs associated with hazardous and "worst forms of child labour."  This concern is often expressed in terms of the need to protect children from hazardous and other worst forms of work, and from all forms of exploitation.
The indicators suggested in table 1 to represent protection of children against hazardous work and exploitation are difficult to measure, and this is reflected in the paucity of quantitative data available for these indicators. This is not surprising. The concept of exploitation in particular is value-laden and therefore difficult to define objectively. However, hazardous and other worst forms of work, although also value-laden since what are considered to be hazardous and worst forms of work vary across cultures and development/income levels, can more readily be defined objectively. With the endorsement by the international community of priority action against hazardous and other worst forms of child labour, as set out in Convention No. 182, it is essential that increased efforts be made to develop ways of measuring these particular forms of child labour.
Development of children
Children develop quickly, acquiring skills and knowledge in preparation for becoming productive adults and citizens. Learning and skills are acquired both through formal schooling (e.g. reading, writing and arithmetic), and through experience of work and life (e.g. self-reliance, responsibility and traditional skills and knowledge).
A major concern about child labour is that it often interferes with children's ability to attend and do well in school. But one must be careful about assuming that all forms of child labour necessarily interfere with school attendance and performance. Though full-time work (whether hazardous or not) is clearly incompatible with school attendance and performance, part-time child labour does not necessarily interfere with them when it occurs during the school holidays, or for a few hours a week during the school year. Although it is uncertain the number of hours children may work during the school year before their school performance suffers, it is likely to be at least 2-3 hours per day, or 15 or so hours per week.  It is also important to bear in mind that learning in school depends greatly on how good the school is.
Although the fact is usually ignored by programmes and policies concerned with the elimination of child labour, valuable skills and knowledge can be learned through work (Boyden, Ling and Myers, 1998). It is also important to recognize that school itself can sometimes be the cause of child labour -- either because children need to earn money to help pay for school costs, or because they see school in a negative light, perhaps because of violence against them at school (Boyden, Ling and Myers, 1998).
The indicators listed in table 1 to represent children's development and learning are poorly measured at present and proxy indicators are generally used. For school-based learning, it is almost always assumed that learning is equivalent to the number of school standards a child has completed. This is unfortunate, first, because schools in some countries are of such poor quality that many children do not learn much there; and second, because this ignores the possibility that light work and non-hazardous work can contribute to learning life skills. Indeed, when faced with a choice between non-hazardous work, poor-quality schools or idleness, families and children might be rational in concluding that non-hazardous work is in the child's best interest, as compared to school attendance or idleness. Third, children's total number of work-hours is rarely measured. In addition, time spent by children on housework and childcare is not considered as labour force activity according to the international definition of th e labour force and therefore by definition is not child labour; yet many girls spend long hours on housework and childcare, and as a result do not attend school. They are as much at a disadvantage in their ability to attend and perform well at school as boys who are full-time wage earners. Fourth, little consideration is given to when in the school year work is done. Yet it is the total number of work-hours performed during the school session which is important in determining possible conflict with school attendance and performance. Some of these measurement issues are discussed below.
The economic and labour market impacts of child labour
A number of important economic and labour market effects associated with child labour should concern policy-makers. These are divided into those which occur at the micro family level, and those at the macro and the meso labour market and economy levels (see table 1):
* micro family level:
-- family income and survival;
* macro labour market and economic levels:
-- labour markets (e.g. wage rates and adult unemployment);
-- economic growth and economic development.
Children's labour is an important source of income for poor families. It is widely believed that poverty is the main (though not the only) reason for child labour in poor countries, and that the survival of many poor families depends on the cash and in-kind income generated by children.
At the micro family level, the economic concern is with the difficulties which poor families and poor children would face in the short run if child labour were eliminated. This implies that child labour programmes should consider poor families' income needs and the consequences for family survival if their child members stopped work completely. It suggests, for example, the usefulness of targeted income transfers and/or subsidies for poor families with children in school; of adjusting school calendars to enable children to work in peak seasons and part time, if necessary; and of providing income-generating opportunities for adult men and women as a substitute for child labour. This also implies that quality schools are essential to encourage and justify the family sacrifices required to eliminate child labour. More research and analysis are needed on the determinants of hazardous and nonhazardous child labour, including on the role of poverty and the identification of situations in which poverty does not pre clude the elimination of child labour; the role of employers and the demand for child labour; and the role played by children's and parents' perceptions of school quality and labour market opportunities.
Turning now to macroeconomic and labour market concerns, there is a general belief …