Explaining Differences Within and Between Countries in the Risk of Income Poverty and Severe Material Deprivation: Comparing Single and Multilevel Analyses

Abstract

This study investigates the differences between the risk factors of income poverty and severe material deprivation by assessing to what extent both indicators are subject to the same underlying household-level and country-level determinants. Given that these two indicators encompass the majority of the Europe 2020 social inclusion target group, it is crucial for policy makers to better know which determinants are effective in explaining the differences within and between countries in the risk of these two target components. We employ a series of single-level and multilevel logistic multinomial models to explain differences between those who suffer from income poverty ‘only’, from severe material deprivation ‘only’, from both problems and none. The comparative use of single-level and multilevel methods allow for the first time to confront the respective within and between-country explanatory power of both types of models. In addition to the usual econometric approach of identifying significant relationships, we employed the Shapley decomposition method to compare the relative contribution of independent variables at the household-level (i.e. work intensity, household education, household costs, socio-demographic variables) and country-level (i.e. size of social spending (total, in-cash, in-kind), pro-poorness of social benefits, median income levels and the unemployment rate) to within and between-country explained variance measures. We apply our method to the cross-sectional European Union Statistics on Income and Living Conditions for the year 2012.