Household Economic Survey (Income) 2022/23

Methodology Information on New Zealand households’ income, housing costs and material well-being is based on data collected by Household Economic Survey (HES) 2022/23. The 2022/2023 HES also included the expenditure components that collected information on household expenditure. It runs every three years. Changes made to HES questionnaire in 2022/23 Only minor changes were made to the survey we collected data in HES 2022/23 compared with HES 2021/22: Phone interviewing was deployed as an option for completing the survey in a COVID-19 restricted environment. After the end of COVID-19 restrictions, it was kept as an option and will continue to be utilized. We now collect contact details from the respondent for conducting both face-to-face and remote surveys, including phone numbers and/or emails. The mortgages module was re-designed to simplify and reduce the amount of questions asked. The output from the mortgages module remained unchanged but how this information was calculated (derived) changed to reflect the changes to the questionnaire. Reference and collection period We collect HES data over the course of a year, from 1 July to 30 June. At the interview, the respondent is asked about their income in the previous 12 months. For example, an interview in November 2022 would collect the household’s income and wellbeing in the 12 months from November 2021. This means that households interviewed for the HES 2022/23 survey in 2022 will include some 2021 income, while those interviewed in 2023 will include some 2022 income. Sample design The HES uses a stratified, multi-stage, cluster design. Primary sampling units (PSUs) are selected from the household survey frame, then dwellings within PSUs, then eligible persons within selected dwellings. A new sample of PSUs was again selected for the 2022/23 HES from the broader sample, which included some overlap with the PSUs selected for 2021/22. A sample size of 20,000 responding households is required to meet the accuracy objectives related to the Act. With 2,500 PSUs, and just over 11.4 households selected per PSU, we get a total selected sample between 28,000 and 28,500 households. We assume that at least 70 percent of households will provide a full response, leaving a final (achieved) sample of at least 20,000 households. (The target sample sizes for the expenditure and net worth components, which are subsamples of the core HES, are 5,500 and 8,500 households, respectively). With ongoing collection challenges, the target achieved sample size for HES 2022/2023 was reduced from 20,000 to 15,000 households. The subsample for HES expenditure was not reduced. This was implemented by removing half of the PSUs allocated to each month from January to July, with the remaining PSUs in the sample retaining the principles and assumptions of the initial sample design. Unbiased collection of the 15,000 households would be expected to achieve the same population and concept representation as the full 20,000 households, but with higher variance and sample error. The total selected sample included 21,100 households. This approach differed slightly to sample reductions in the previous three years, which were disrupted by COVID-19 alert level restrictions and involved removing whole months of allocated sample. The intention was to minimise disruption to the intended design for HES expenditure, which comprises a subsample of the core HES and is also allocated to each month of the collection period. There was no reduction to the target sample of 5,500 households for HES expenditure. Impact on data quality The sample reduction and the smaller achieved sample than designed for have potential implications for data quality in HES 2022/2023. A similar approach to one used in the 2021/2022 HES is employed to assess the HES 2022/2023 data quality. Impacts of disrupted data collection on 2022 Household Economic Survey statistics | Stats NZ details the full assessment approach, including how we employed the Stats NZ Statistical Quality Model and the Total Survey Error (TSE) Framework. However, users should be aware that the reduction in sample size means that the statistics are subject to higher sampling error. Caution is advised in interpreting statistics for subpopulations, where the sampling error and risk of bias is higher. Data input Assignment of cases is centralised in Salesforce, a system that allows for real-time observation of response rates. The team of interviewers use BLAISE to conduct household surveys. BLAISE is a computer assisted interviewing (CAI) software that guides the interviewer through the correct sequence of questions. They are displayed one at a time and have automatic routing built in to make sure that respondents are only asked questions that are relevant to them. Once submitted, the data is stored and a response is logged in our centralised tracking software, Salesforce. This allows for a real-time overview of progress in reaching the target response rates for demographic groups. The data is then fed through various editing stages, before being loaded into the processing database, EPIC. Despite our best efforts to obtain accurate information about respondents’ income, relying on survey responses for data of this nature inevitably introduces uncertainty. For example, respondents may not remember or may fail to disclose all sources of income over the past year to the interviewer. They may provide only ‘rough estimates’, describe income after tax, or forget changes to their regular income over the year. In some cases, family members may not know the income of other family members. Benefit income is often understated – as people can forget benefit income when it was received only for small periods throughout the year. We combine survey data on income with admin data from the Integrated Data Infrastructure (IDI), a large research database managed by Stats NZ holding microdata about people and households. It contains full tax information related to individuals, including data provided by employers for each employee (the employee monthly schedule), self-employment income, and some investment income. We also use data provided by the Ministry of Social Development (MSD) about benefits paid, including Working for Families (WFF) tax credits, and accommodation supplement. Since HES 2018/19 we have used admin data to provide annual salary and wage and government transfer income. In HES 2018/19, we asked respondents their income but used the admin data in the published statistics. Starting with HES 2019/20, respondents have not been asked to provide their income amounts for these income variables. Some income sources are not currently available in the IDI, including investment income, some sources of irregular income, and non-taxable income. We collect these income variables directly from respondents. Salary and wage income is provided on an individual’s pay day and is updated in the IDI on a quarterly basis. However, other income (for example, from self-employment) relies on individuals providing their tax returns, which may be delayed before being included in the IDI. Due to this timeliness issue, we use the self-employment income provided to us by the respondent. Information on income received from the WFF scheme is available from IRD and from MSD and this data is used for relevant households. However, for some households this income is received annually and there can be delays in this information being incorporated in the IDI due to delays in filing tax returns. For this reason, annual income from WFF is estimated for some families and is revised in the following year when more information is available. Linking admin data The use of admin data in HES requires linking individuals in HES to the IDI spine, a secure dataset to which all datasets in the IDI are linked. This link uses address, address history, name, and date of birth. The ‘link rate’ refers to the proportion of individuals in the HES sample that are successfully linked to the IDI. A high link rate ensures high quality data. The link rate of the overall HES sample in October 2022 to the IDI was 95.2 percent (with a false positive rate of 1.8 percent). The link rate for children is lower than for adults because we do not collect date of birth for children in the HES (although we do ask about age), and there is therefore less information that can be used to successfully link to the IDI. However, this does not affect the child poverty measures because these are determined from household income, which is based on adults in the household. Linking to the IDI enables us to assign admin income data (salaries and wages, and benefits) to all in-scope and eligible individuals aged 15+ from responding households, even if they themselves did not respond to the survey. This increases the number of usable responses in the dataset. Records unable to be linked to the IDI have wage and salary and benefit income imputed to reduce potential bias. Imputation for HES 2022/23 Imputation in HES replaces missing values with actual values from similar respondents. For HES 2022/23, we imputed missing values for the following variables: Income o employment earnings and government transfers where a respondent has not been linked to the IDI o self-employment income where respondent is known to have such but has not provided a value o investment income where respondent is known to have such but has not provided a value. Housing costs o local and regional authority property rates for primary property Person demographics o age o gender o sex at birth o ethnicity o disability status for people over the age of 2 years Response rate for HES 2022/23 We set a target achieved sample rate of 70% and aim to achieve at least 20,000 responding households from 28,500 households that are initially selected. However, for the 2022/2023 HES, the achieved target sample size was reduced from 20,000 down to 15,000 households. The final achieved sample in the 2022/2023 HES included 14,100 households (overall achieved sample rate of 67 percent). The achieved sample rate (ASR) is calculated as the number of eligible households that responded divided by the total number of dwellings sampled. Essentially, it tells us what percentage of the sample responded to the survey. Eligible responding sample ASR = ____________________ Ineligible sample + eligible responding sample + eligible nonresponding sample The response rate for HES 2022/23 was 76 percent. Response rate is also monitored at regional level and by NZDEP2018 status. The response rate is effectively an estimate of the proportion of the whole population that would have responded, from those that were eligible. It is calculated as the weighted number of eligible households that responded to the survey divided by the estimated number of total eligible households in the population. Eligible responding population Response rate = ___________________ Eligible responding population + eligible nonresponding population Besides being calculated using the sample, rather than the estimated population, the ASR differs from the response rate because it includes the ineligible dwellings in the denominator. This difference means that the ASR is less sensitive to the classification of household eligibility and is more stable over time than the response rate. Sample errors Sample error is a measure of the variability that occurs by chance because a sample rather than an entire population is surveyed. We can calculate the level of uncertainty around a survey estimate by exploring how that estimate would change if we were to draw many survey samples for the same time period instead of just one. This allows us to define a range around the estimate (known as a “confidence interval”) and to state how likely it is that the real value that the survey is trying to measure lies within that range. Confidence intervals are typically set up so that we can be 95% sure that the true value lies within the range – in which case this range is referred to as a “95% confidence interval”. We calculate sample errors using the jackknife method which is based on the variation between estimates of different subsamples taken from the whole sample. The tables below summarise the sample errors between 2018/19 and 2022/23 by income source and housing-cost type. The tables also indicate the variability of the estimates between the five surveys. Customers should take care when interpreting income or housing-costs estimates with sample errors greater than 20 percent – they are statistically less reliable than estimates with sample errors less than or equal to 20 percent. Sampling errors for average annual household income, by income source (over all households) Year ended 30 June 2019-2023 Income source Level sampling error (%) 2018/19 2019/20 2020/21 2021/22(R) 2022/23 Wages and salaries 1.5 1.6 1.5 2.0 1.6 Self-employment 8.7 7.0 5.5 7.4 7.4 Investments 11.4 6.8 6.2 9.8 18.5 New Zealand Superannuation and war pensions 1.2 1.0 1.3 1.4 1.1 Other government benefits 2.3 2.8 2.2 3.7 2.5 Other regular sources 6.9 7.2 7.8 9.5 10.6 Total Gross income 1.6 1.3 1.1 1.5 1.6 R Revised Sample errors for average weekly household expenditure, by housing cost type (for households with that type of expenditure) Year ended 30 June, 2019 -2023 Expenditure item Level sampling error (%) 2018/19 2019/20 2020/21 2021/22(R) 2022/23 Property and ground rent 1.9 1.8 1.9 2.5 2.0 Other payments connected with renting 8.2 11.3 6.9 9.4 7.8 Total rent payments 2.0 2.0 2.0 2.6 2.1 Mortgage principal repayments 2.6 2.6 2.6 3.3 6.3 Mortgage interest payments 3.6 2.9 3.3 4.1 3.7 Application and service fees for mortgages 16.8 20.8 21.4 29.0 32.4 Total mortgage payments 2.3 2.3 2.4 2.8 3.9 Property rates 1.5 1.8 1.6 2.5 1.8 Building related insurance 2.1 1.7 1.7 2.0 3.7 Other housing costs 10.8 21.2 13.6 17.1 16.6 Total housing costs 1.7 1.6 1.9 1.9 2.2 R Revised External influences Changes in income and housing costs may be influenced by one-off real-world events. Events that could have influenced the HES 2022/23 data are: the Consumer Price Index (CPI) annual inflation rate reached a peak of 7.2 in June 2022, then slightly decreased to 6.0 in June 2023. to combat inflation, the Reserve Bank’s Monetary Policy Committee have been increasing the Official Cash Rate throughout the reference period. The remit of the Monetary Policy Committee is to keep inflation between 1% - 3%. The timeline for increases is as follows: o 13th July 2022, the OCR increased from 2.0 to 2.5. o 17th August 2022, increased to 3.0. o 5th October 2022, increased to 3.5. o 23rd November 2022, increased to 4.25. o 22nd February 2023, increased to 4.75. o 5th April 2023, increased to 5.25. o On the 24th of May 2023, the OCR is 5.5. Higher prices for interest payments are the biggest contributors to the 7.2 percent increase in the cost of living for the average household (HLPIs) in the 12 months to June 2023 minimum wage had two increases during the reference period, the first being from 1 April 2022 from $20.00 to $21.20, and the second from 1 April 2022 from $from $21.20 to $22.70. In our data there will be full capture of the 2022 increase, and partial capture of the 2023 increase. increases in the starting out and training wage from $16.00 to $16.96 from 1 April 2022 and then to $18.16 effective 1 April 2023 changes to benefit system as part of Budget 2021 that came into effect on 1 April 2022 o Jobseeker Support, Supported Living Payment, Sole Parent Support, and other main benefits increased 4.71% in line with average wage increase. o NZ Super, Veteran’s Pension, Student Allowance. Orphan’s and Unsupported Child’s benefit and other supplementary assistance increased 5.95% in line with the Consumers Price Index o all main benefit rates also increased by $15 per week for families with children o Student Allowance increased by $25 per adult, per week. o paid parental leave increased from to $621.76 per week to $661.12. o Working for Families tax credits increased o Orphans Benefit and Unsupported Childs Benefit increased by between 12 and 26 percent per week. The 2022 Budget introduced a cost of living payment of $350 that was paid to individuals aged 18 or over, earning $70,000 or less, who were New Zealand tax residents living in New Zealand and weren’t eligible for the Winter Energy Payment. The 2022 Budget also included other measures such as extending the reduction in fuel excise duty and road user charges, and public transport kept at half-price. changes to benefits for 2023 financial year included the following: o Jobseeker Support, Supported Living Payment, Sole Parent Support, and other main benefits increased 7.24% in line with average wage increase of 6.24% plus an extra 0.98 percentage points to that the overall increase matched the increase in the Consumer Price Index. o NZ Super, Veteran’s Pension, Student Allowance. Orphan’s and Unsupported Child’s benefit and other supplementary assistance increased 7.22% in line with the Consumers Price Index o Student Allowance payments increased in line with inflation, with single students under 24 without children to get an extra $20.21 per week o Childcare Assistance income thresholds increased. o Working for Families tax credits increased, including an extra $4 for Best Start Payments taking it to $69 per week and an increase of $9 for the eldest child rate of Family Tax Credit lifting it to $136 per week. the labour cost index has been increasing. For the year to the March 2023 quarter, it increased 4.3%, (unadjusted 5.8%) unemployment remains at record lows for the reference period of HES 2021/22. In the June 2022 quarter unemployment was at 3.3%, and remain stable for September, December, and March quarter, it rose slightly to 3.6% for the June 2023 quarter. en-NZ