Formats and related files
Staff and teams are writing in their individual capacity and the views in this paper are not necessarily a “Treasury” view. Please read Disclaimer for research and commentary publications.
This work makes use of Stats NZ's Integrated Data Infrastructure (IDI), please also read the IDI disclaimer.
The code used to produce the detailed results in this report can be accessed at the following GitHub address: https://github.com/Treasury-Analytics-and-Insights/analytical-note-22-04-insights-from-New-Zealand-child-poverty-data
This file presents the results using a clustering algorithm:
This second file presents similar results but for groups based on the overlaps of three poverty indicators:
Extract from paper#
One way that governments support people is by providing a safety net through main benefits like the Job Seekers Allowance, supplementary benefits like the Working for Families tax credits, and discretionary payments such as special needs grants.
A central goal of these programmes is to reduce the number of families below a minimum standard of living – in other words to reduce the number of people in poverty. But while this may be a simple idea, in practice it is no easy task.
One challenge is that there is no single, objective measure of what it means to be poor. Indeed, it has been said that “counting the poor is an exercise in the art of the possible” (Stephens & Waldegrave, 2001), where the “art” lies in choosing a poverty indicator. The best approach is to use a range of poverty indicators that illustrate different parts of the puzzle and together provide a fuller picture, enabling others to make their own judgements.
This analytical note outlines an approach that uses the available data to provide insights into three different indicators of poverty, making use of recent data and modelling improvements. It applies a statistical algorithm to identify seven different categories of children in poverty and describes the characteristics of children in each group.
This exploratory analysis confirms that the relationship between material hardship, income, and housing costs is complex. For some of the identified categories there is a direct relationship between low incomes, either before or after housing costs, and material deprivation. However, for several categories low incomes do not correspond to deprivation and vice-versa.
- There are strong links between hardship, income, and housing costs for some families, but not others. For beneficiary families, we find that families experiencing deeper income poverty are more likely to be experiencing hardship. But there is less of a link between hardship and income when we consider working families. Many appear to have adequate incomes but experience deprivation and vice versa.
- The data suggest that we also need to think about other aspects of economic wellbeing such as wealth and extra costs related to, for example, disability and childcare.
- Not all beneficiaries are in poverty, and not all children in poverty are in beneficiary families.
- Beneficiaries with high housing costs have their before housing cost incomes boosted via the Accommodation Supplement, which makes them appear to have adequate incomes even though they are in poverty on other measures.
- However, for working families with high housing costs, our model suggests that some families in poverty meet the eligibility requirements for Accommodation Supplement but are not receiving it.
- Although mostly coupled parents, working families in poverty or near poverty thresholds are more likely to be one earner families; these families are around twice as likely to have only one earner than other families with children.