Child development and family income in the United Kingdom: does money matter?

Presenter: Mara Violato, University of Oxford

Abstract

Recent empirical literature, mainly North American, indicates that children born into poorer families are at greater risk of poorer cognitive, behavioural and health outcomes than their wealthier peers. These poor outcomes may also adversely affect those children’s future educational achievements and labour market engagement. If the observed positive association between child outcomes and family income is causal, government policy interventions such as money transfers to assist low income families with children might be an effective way to reduce economically-driven health inequalities in childhood.

Estimating the effect of family income on child outcomes using observational data is challenging. Unobserved data heterogeneity makes the identification of causal relationships - so crucial in policy making - problematic. This paper contributes to the existing empirical literature by attempting to overcome the potential endogeneity of family income using three econometric approaches: a ‘mopping up’ technique, an ‘instrumental variable’ approach and a fixed effects model.

We analyse data from the Millennium Cohort Study, a nationally representative birth cohort study collecting information on health, wealth, education, employment, family characteristics and lifestyle of almost 19,000 children born in the United Kingdom in 2000-2001. We concentrate on one measure of child cognitive development, the British Ability Scale Naming Vocabulary, and one measure of child behaviour, the Total Difficulties Score. Both child outcomes were measured when the child was aged three and five. Our models are specified according to the insights provided by the ‘family investment’ and the ‘family stress’ theories, which underlie the child cognitive outcome – family income relationship. They therefore take into account socio-economic factors, parents’ general and mental health, parenting practices and parents’ monetary and non-monetary investments in children’s development. Detailed information on fathers’ role, which has largely been neglected by previous literature, is also included in our models.

Overall, our preliminary findings suggest that income has a causal effect on child development, but the magnitude of this impact is quite small. From a policy perspective this suggests that the use of income redistribution as a tool for ameliorating child development may not be as cost-effective as it is often thought. It might also indicate that a service-oriented strategy would be more cost-effective and politically feasible and could, therefore, complement income transfer strategies. We argue that our evidence could play a relevant role in informing a more efficient and equitable allocation of scarce resources available to governments to tackle child health inequalities.