This study assesses the impact of income and asset poverty on child work using the rural sub-sample of the 2004 Malawi Integrated Household Survey. Instrumenting consumption expenditure with a location dummy variable and interacting consumption expenditure with household land-holding size in probit models, the likelihood of child labour is found to relate negatively with household consumption. On the other hand child labour relates positively with household land-holding size for consumption poor households only and when labour markets are imperfect. These findings do not discourage asset accumulation policies as a remedy against child labour but support policies that aim at increasing returns on the assets.