Table 1. Overview of key differences in how in-work poverty is measured internationally Bureau of Labor Statistics (USA) Eurostat (EU) ILO Definition of Working Activity Status Working or looking for work Worker Worker Amount of work >6 months per year >6 months per year At least 1 hour per week Income before taxes and tax credits Anchored (US Official Poverty Line) Net income equivalised by ‘modified’ OECD scale Per capita expenditure Relative (60% median income) Absolute ($1.25/day) Definition of Poverty Income Concept Thresholds Source: Lohmann 2018. Lohmann (2018) does point to one important commonality between the definitions presented in Table 1: he notes that “all define an in-work poor person as a working person who lives in a poor household” (2018, p. 7). This results in the two concepts of work and poverty having inconsistent units of analysis (that is, the individual versus the household). To show the significance of this inconsistency, it is worth illustrating this measurement issue with a specific example. Imagine a family or household comprises one man, one woman and two dependent children. The woman is in employment and the man is not; the children, aged three and seven, are too young to work. The family is living in poverty. In the approaches presented in Table 1 (regardless of definition adopted), only the woman would be counted as experiencing in-work poverty, since only she is working. Her partner and their children are dropped from the enquiry since they do not work. When it comes to thinking about the policy responses to tackle in-work poverty, we are implicitly pushed towards thinking about how to address her poverty, rather than trying to assess and ensure income adequacy for the whole household. Proposals to tackle in-work poverty prevalence are therefore likely to focus on her level of pay (hence why the debates regarding in-work poverty and low pay are often conflated) or perhaps on encouraging her to undertake additional hours of paid work. The former is particularly problematic as the international literature demonstrates with great consistency that the relationship between low pay and in-work poverty is weaker than assumed (Marx et al., 2012; Maȋtre et al., 2018; see also Bennett, 2014 for a discussion). The strength of the relationship depends crucially on whether there are other income sources in the household in addition to the earnings of the low paid member and how extensive the household’s needs are. There are other ways the hypothetical household described above may escape poverty besides options focussed solely on the female worker. For instance, her partner might return to work so that there would be two earners in the household. Or, if he were unable to do so because he is taking care of the younger child, the government could help via increasing access to childcare. In short, by focussing solely on the individual worker, we run the risk of ignoring the fact that policy solutions focussed on the non-working members might help the household escape poverty. Examples of studies that do employ consistent units of analysis in relation to the work and poverty variables are few and far between. Of note are a few recent UK studies (Tinson et al., 2016; McInnes et al., 2013; Hick & Lanau, 2017) that focus on all working-age adults living in households (who meet the relevant definition of experiencing poverty and having someone who is in employment), not just the workers. Further, as will be shown in Section 3, our approach also utilises consistent units of Page 12