Child poverty statistics released

Child poverty statistics released today will be used as baseline rates for the purposes of the Child Poverty Reduction Act 2018, Stats NZ said.

The Act was introduced to help achieve a significant and sustained reduction in child poverty. The Act requires the Government Statistician to report annually on 10 measures, with the first reporting being for the year ended June 2019.

There is no clear line between poor and non-poor people, as all poverty measures require judgement calls that reflect underlying assumptions. However, the range of threshold lines reported in this release have general support.

“Together the nine measures reported in today’s release provide a comprehensive picture of child poverty in New Zealand that one measure alone cannot provide,” household statistics senior manager Jason Attewell said.

We have made improvements to the data source for measuring child poverty in the future, including increasing the sample size from between 3,000 and 5,500 to 20,000 households. These improvements were implemented in the 2018/19 HES, which is collected between July 2018 and June 2019. First child poverty results from this survey will be available early in 2020.

“In the meantime, we used additional data sources and methodologies to produce estimates for the year ended June 2018 that are as robust as possible,” Mr Attewell said.

“This ensures that the baseline estimates released today are robust enough to be used by Government for setting 3- and 10-year targets for reduction of child poverty for three primary measures specified in the Act.”

Data and methodology caution provides more information about the additional data sources and methodology we have used and why the estimates should be used with caution.

Low income before housing costs are deducted

In the year ended June 2018, 16 percent of children lived in households with less than 50 percent of the median equivalised disposable household income before housing costs are deducted. While there has been some variability over the last five years, the trend has been fairly flat.

“Equivalised household income means that we adjust for the number of adults and children in a household. We do that because as the size of a household increases, more income is required to maintain the same standard of living,” Mr Attewell said.

This measure compares a household’s income for the previous 12 months to the current median for all households. The median will move from year to year due to inflation and economic changes. A low-income household will improve its situation if its income moves by more than the median.

Figure 1

Note: Error bars in figure 1 show 95 percent confidence intervals, between which we are confident that the true rate lies.

Text alternative for figure 1, Percentage of children living in households with less than 50 percent median equivalised disposable household income before housing costs are deducted, 2013–18.

Income after housing costs are deducted – fixed-line measure

Looking at income before housing costs is one way of looking at child poverty; another is to look at income after housing costs are deducted. This measure gives an indication of the spending power households have after they’ve paid housing costs, such as rent, mortgage repayments, rates, and insurance.

The ‘after housing costs’ measure compares a household’s income in the current year with the median income for all households in the 2017/18 year adjusted for inflation. This allows us to understand if a household’s real inflation-adjusted income has improved irrespective of what happens to the incomes of other households.

In the year ended June 2018, 23 percent of children lived in households with less than 50 percent of the median equivalised disposable household income after housing costs are deducted. This rate has been gradually decreasing since the year ended June 2015.

Figure 2

Note: Error bars in figure 2 show 95 percent confidence intervals, between which we are confident that the true rate lies.

Text alternative for figure 2, Percentage of children living in households with less than 50 percent median equivalised disposable household income after housing costs are deducted (for the 2017/18 base financial year), 2013–18.

Material hardship

In the year ended June 2018, 13 percent of children lived in a household experiencing material hardship, which indicates that they are missing out on more than a handful of things that could be expected in a typical household. It is important to note the relatively large sample error that remains on this estimate.

“Material hardship is a way of measuring poverty by understanding what households don’t do or don’t have because of financial constraints. Examples include not having two pair of good shoes, putting off a visit to the doctor, or not being able to pay the gas or electricity bills on time,” Mr Attewell said.

Figure 3

Note: Error bars in figure 3 show 95 percent confidence intervals, between which we are confident that the true rate lies.

Text alternative for figure 3, Percentage of children living in households experiencing material hardship, 2013–18.

Child poverty statistics: Year ended June 2018 contains Excel tables that report on the six additional measures required by the Act.

From July 2025, we will report on the poverty persistence measure, which looks at the length of time a person has been in poverty.

The estimates released today are only at a national level, given the limitations of the relatively small sample size.

In early 2020 we plan to publish estimates at a regional level and by ethnicity, based on results from the larger HES of about 20,000 households.

Data and methodology caution

The estimates in today’s release are based on household economic surveys (HES) from 2012/13 to 2017/18. HES is a random sample survey of between 3,000 and 5,500 households, of which around a third are households with dependent children. It is well suited for and delivers valuable information for purposes such as the overall distribution of household income and material wellbeing, and relativities between groups. However, when the focus is on very short-term changes, especially year-on-year or when more precision is required in a given year, the current HES is not able to deliver robust results given its relatively small sample size.

We also tend to have lower response rates from households in low socio-economic areas, which means that these households are often underrepresented in the sample.

We have made improvements to the data source for measuring child poverty in the future, including increasing the sample size to 20,000 households. These improvements were implemented in the 2018/19 HES, which is collected between July 2018 and June 2019. First child poverty results from this survey will be available early 2020.

In the meantime, we partially addressed the limitations in the survey for the current release with the use of different data sources and methodology. Using the Integrated Data Infrastructure (IDI) enabled us to increase the effective sample size by combining income data from admin sources with HES and household labour force survey data. We used this for the ‘low-income before deducting housing costs’ measures.

There is no admin data available on either housing costs or material wellbeing. Therefore, we had to use an alternative approach for ‘after housing cost income’ and ‘material hardship’ measures. For these measures, we partially addressed the underrepresentation of households from low socio-economic areas through reweighting the sample.

This ensured that the child poverty estimates released for 2017/18 are robust enough to use as baselines for setting 3- and 10-year targets.

Text alternative for figure 1

Figure 1 is a line graph with error bars showing estimates on the percentage of children living in households with less than 50 percent median equivalised disposable household income before housing costs are deducted, with sample error bars for the financial years 2013 to 2018. The graph shows that while there has been some variability over the last five years, the trend has been fairly flat. The figure shows estimates between 14 and 16.5 percent, with the latest estimate being 16.5 percent, and the rate gradually decreasing since 2014/15. Note: the sample error on this latest estimate is 1.1 percentage point.

Text alternative for figure 2

Figure 2 is a line graph with error bars showing estimates on the percentage of children living in households with less than 50 percent median equivalised disposable household income after housing costs are deducted (for the 2017/18 base financial year) for the financial years 2013 to 2018. The graph shows estimates between 22.4 and 30.9 percent, with the latest estimate being 22.8 percent, and the rate gradually decreasing since 2014/15. Note: the sample error on this latest estimate is 1.9 percentage point.

Text alternative for figure 3

Figure 3 is a line graph with error bars showing estimates on the percentage of children living in households experiencing material hardship (have a DEP-17 score of six or more) for the financial years 2013 to 2018. The graph shows estimates between 12.1 and 18.1 percent over the years 2013 to 2018, with the latest estimate being 13.3 percent. Note: the sample error on this latest estimate is 2.3 percentage point.