A distributional decomposition of birthweight differences by maternal education: A comparison of France and the UK

While socio-economic gradients in newborn health are well established (Currie, Am. Econ. Rev. 101(3), 1–22, 2011), understanding what produces these inequalities, and in particular the unique contribution of inter-connected mechanisms, remains difficult to estimate. We adapt decomposition methods proposed by Rothe (J. Bus. Econ. Stat. 33(3), 323–337, 2015), which isolate the marginal contribution of potential explanatory variables, to explore differences in birthweight by maternal education, in France and the UK. These methods allow looking at the whole birthweight distribution rather than binary indicators of low birthweight, and the adaptation we propose allows grouping covariates into sets that tap into the same concept, providing a better measurement of the overall role of each potential mechanism in explaining health gaps. Analyses are applied to rich, nationally representative data for infants in France and the UK. Beyond differences in birthweight distributions by maternal education, differences in group characteristics (health behaviours during the pregnancy; maternal work condition; social support; demographic and economic factors) explain much of the gaps in birthweight across educational groups in both countries, but especially France. Smoking appears to be the largest source of inequality in both countries. However, differences between countries emerge, such as the importance of social support and mental well-being in the UK, suggesting that national contexts may matter in generating differences in birth health across education groups.