Working Paper No. 2023-29 - Changing Household Structures, Household Employment, and

8 The intuition behind these counterfactuals is that comparing the observed poverty rate in 2019 with the first counterfactual rate tells us the effect of the change in the distribution of household types between 1986 and 2019 on poverty, and thus its contribution to the observed change in poverty over the period. [...] Changing Household Structure/Employment Profiles and Trends in Poverty We now assess the implications of these changes in household structure and employment patterns for the overall poverty rate by asking what the poverty rate in 2019 would have been if each of the household types we distinguish faced its actual poverty risk for that year, but the composition of the sample in terms of those types. [...] By contrast, at the other end of the spectrum Denmark and the USA saw only quite limited changes in the profile of sample households over the period and no increase in the share of two-earner couple households, which is why the counterfactual exercise shows such little impact on the aggregate poverty rate. [...] A potentially important aspect of the measured impact of the changes in profile is that these affect the mean income of the sample and thus the (half mean) poverty threshold being applied in the counterfactual. [...] Changes in the composition of the sample by household type had only a modest impact in the UK and made very little difference to the trend in poverty in Denmark, Hungary, and the USA.