1. INTRODUCTION “Is Work the Best Antidote to Poverty?” (OECD, 2009, p. 165). This question was the title of an OECD book chapter regarding in-work poverty across the OECD. The analysis that followed illustrated that while employment does reduce poverty risk, there was still a sizable likelihood of experiencing inwork poverty in many OECD economies. In particular, on average in the OECD area, 7 percent of individuals classified as living in working households were identified as poor (OECD, 2009), with it ranging from 2 percent in Norway to more than 10 percent in countries such as Japan (11 percent), United States (12 percent), Mexico (17 percent) and Turkey (17 percent). New Zealand was found to be close to the OECD average at 8 percent. 4 In general, while awareness of ‘in-work poverty’ has been increasing in recent years in many Western economies, little is known about this phenomenon in New Zealand. Therefore, the aim of this study is to undertake the first explorative step regarding in-work poverty in New Zealand. It seeks to establish the prevalence of it, as well as characterise the affected populations. In-work poverty is often associated with wider debates about the declining fortunes of the middle class, or the emergence of a ‘gig’ economy, or with the rising cost of housing (e.g., NZ Herald, 2018; Guardian, 2018). However, while such debates are prominent in both political life and public discussion more generally, we know less about the actual prevalence of in-work poverty in New Zealand and lack empirical evidence regarding which groups are at greatest risk of experiencing it. The key aim of this study is to fill this evidence gap. Before outlining the empirical approach undertaken, it is worth first outlining a number of key findings from the international literature on this topic. The first is that the association between low pay and in-work poverty is weaker in practice than is often assumed in public debate. As Maȋtre et al. (2018, p. 124) note “low pay for an individual worker may or may not be associated with poverty and disadvantage for the household in which they live”. This, they note, is because total household income, and thus the determination of whether a household is in poverty or not, is based on all income sources, considering earnings from all members of the household, as well as non-employment income, such as social transfers. 5 Previous overseas studies have also found a number of characteristics associated with an elevated rate of in-work poverty. These include being a migrant, working in service sector occupations (Crettaz, 2011), single-parent households, and individuals living in rented accommodation (Hick & Lanau, 2017). To the extent that data permits, we have sought to examine whether some of the trends observed in the international literature are also witnessed in New Zealand. For our study, we make use of linked data both from administrative sources and surveys. The two main sources of information are the 2013 Census (which enables us to identify households and their respective characteristics) and Inland Revenue (which is used to identify the income of the A working household was based on whether “at least one adult has a job, at some point during the year” (OECD, 2009, p. 171); the poverty threshold was set at 50 percent of median equivalised disposable income of the whole population; working-age ranged from 15 to 64. The definitions utilised in this report differs, as shown in Section 3. Therefore, comparisons between the OECD statistic for New Zealand and that obtained in this study cannot be reliably undertaken. 5 According to Eurostat (2016), social transfers “cover the social help given by central, state or local institutional units. They include: old-age (retirement) and survivors’ (widows' and widowers') pensions; unemployment benefits; family-related benefits; sickness and invalidity benefits; education-related benefits; housing allowances; social assistance; other benefits.” 4 Page 9