Cross-national comparison of monetary and multidimensional child poverty in the European Union: puzzling with the few pieces that the EUSILC provides

This paper investigates the degree to which monetary poverty and non-monetary deprivation measures identify different groups of vulnerable children; the degree to which children suffer from multiple deprivations across well-being domains; and whether these findings differ systematically across countries with similar living standards. Using the 2007 wave of the EU-SILC (European Union Statistics on Income and Living Conditions), we compare the European Union 'at-risk-of-poverty' indicator to a range of child deprivation indicators. We analyse 13 deprivation indicators that are grouped under four domains (financial strain, housing, neighbourhood and access to basic services) in four countries (Germany, France, the Netherlands and the UK). We use the union approach as aggregation procedure, considering a child to be poor in a particular domain when they are poor with respect to at least one of the deprivation indicators within that domain. Firstly, we find that the Netherlands most often ends up as best performer, while the UK most often ends up as worst performer, with France and Germany in the middle. Nevertheless, consistency in rankings seems to be the exception rather than the rule, suggesting that stochastic dominance techniques can only provide limited solace. Secondly, while there is no consistent evidence that children are more at risk of deprivation than the population as a whole at an indicator level, this picture changes once aggregation proceeds to a domain level and across domains: children become disproportionately more at risk. Finally, we find that higher levels of double deprivation do not necessarily translate into significantly higher odds of double deprivation - an income-deprived child in the UK is equally likely to be also deprived in another domain as an income-deprived child in the Netherlands (this finding also holds across other domains). Our study thus cautions against the construction of a composite index of child well-being; a more promising way of taking multidimensionality into account would be to develop measures that take the 'breadth' of child deprivations into account (across indicators and/or domains).