Paper short abstract:

What explains the high levels of child labour found in some regions of Kenya? Inequalities behind child labour can be examined using a statistical method. We test for social and regional differences, including household and personal characteristics. Area differences in education are found.

Paper long abstract:

We analyse the reasons for very high levels of child labour found in some regions of Kenya. We approach this by applying multilevel modeling with an evidence-based approach and a multidisciplinary orientation. Our interpretation has multiple explanatory strands. We test for explanations of high/low child labour risk, both with household and personal variables, and variables representing districts' characteristics. The prominence of mining and quarrying in a district offers significant explanatory purchase on Kenya's recorded child labour. A district's poverty level is also a risk factor. We tested socio-cultural hypotheses, combining them with the usual economic hypotheses. We found that Kenyan children living with a more educated female carer are less likely to be in harmful child labour. This paper uses a cautious definition consistent with both ILO and UNICEF definitions, which includes farming work, domestic work, and work in goods and services above set weekly hours-thresholds. We used data from 2015/6 Kenyan Integrated Household Survey.