Alexander Krauss outlines why analysing complex phenomena like child labour needs to be done using cross-disciplinary approaches and mixed methods.
Economists commonly assume that monetary poverty is the most important explanation for child labour – and for some, the only serious explanation. Yet to really understand child labour we need to get beyond the popular but reductionist, theoretical idea in textbook economics of a ‘luxury axiom’, which is that children only work when household income drops very low.
If we instead drop standard economic theory and become open-minded about using cross-disciplinary approaches and combining quantitative with qualitative evidence, we realise that household decisions for or against child labour are rarely ever the consequence of one single monetary factor or event. Rather often, they are – as evidence in the agrarian economy of Ghana illustrates – related to a set of factors and events ranging from the structure of the economy, social norms viewing child labour as part of socialisation, and low economic returns to basic education in rural areas, to low government capacity to enforce anti-child labour laws, low quality education often deterring some children from school, the seasonal demand for work on the family farm, and demographic variables such as low parental education. Just taking monetary poverty as a point of departure can therefore constrain a more holistic, cross-disciplinary understanding of why both poor as well as non-poor children work in many countries (with about half of all child labourers in Ghana living above the national poverty line and with about 7 per cent of all children working who live in the richest 40 per cent of households).
On the empirical side, most studies do not combine household survey analysis with qualitative data analysis such as interviews with government officials and children, and therefore they often ignore critical factors such as cultural norms or attitudes about child labour and the institutional capacity to deliver good quality education that are generally not well captured in household surveys. On the theoretical side, the discussion on child labour is driven by models that at times implicitly claim to identify universal ‘truths’ as they often – without any empirical analysis – already contain the ‘causes’ of why children work and the preset policy responses. They often neglect the idiosyncratic and heterogeneous sources of child labour, i.e. the specific traits of different groups of children of different gender and age involved in different economic activities in different geographic areas within different countries with different norms and institutions over different seasons and time periods. Therefore, while a strong interrelationship between household poverty and child labour appears intuitive and is argued and modelled theoretically at the global level, neither such intuition nor such universal theories seem to hold strongly with quantitative or qualitative evidence in agrarian economies like Ghana.
The results of this quantitative and qualitative analysis suggest that for the government to simply focus on poverty reduction, economic growth or school participation will likely not be sufficient in tackling child labour. This is because many child labourers live in non-poor, land-owning households, because growth has often been stronger in manufacturing and services than in agriculture, and because most child labourers already combine work with school. This simplistic uni-disciplinary view is why hundreds of theoretical and empirical articles are often left without viable answers when analysing why levels of child labour can stagnate in poverty-reducing economies, growing economies or countries with expanding levels of education like Ghana. The methodological shortcomings of the standard economic thinking on child labour are however not unexpected as the majority of studies in this literature are conducted by economists who tend to only analyse quantitative data and just focus on the specific subset of economic sources of child labour.
At the same time, the quantitative analysis of complex human phenomena like child labour faces important measurement constraints. An example is that the standard approach in the literature of creating variables for whether children are working or not or whether they are in school or not neglects the complexities of these dynamic phenomena with their many shades of grey. Such variables – which are therefore made mathematically precise and amendable for statistical analysis – cannot always take into consideration if children enrol but never actually attend school; only attend school or work one day a month or only for an hour a day; if they currently do not attend school because their teacher is absent or their school is temporarily closed; or if they currently do not work because it is not a planting or harvesting period for their family’s particular crops.
How precisely we define our variables and make them measurable is restricted by the specific data that can be made available, which predetermines the specific type and scope of possible correlational or ‘causal’ claims that can be made about child labour – or any phenomenon, for that matter. Generating correlational or ‘causal’ conclusions here is therefore limited to those aspects of human phenomena that we can force into simplified, quantifiable categories and make claims about their probability of occurrence.
While any analysis suffers from theoretical, methodological and empirical assumptions and limitations, researchers studying child labour need to better combine different perspectives and methods from across the economic, social and behavioural sciences. This has important implications for the typical way most economists view and try to mitigate harmful forms of child labour that conflict with children’s education and later opportunities, with a focus on monetary income not always leading to the best designed and targeted policies.
Read the full paper by Alexander Krauss: Understanding child labour beyond the standard economic assumption of monetary poverty. Cambridge Journal of Economics, Oxford University Press.
Alexander Krauss is a Visiting Research Fellow at LSE. He also teaches at UCL.
The views expressed in this post are those of the author and in no way reflect those of the Africa at LSE blog or the London School of Economics and Political Science.
