Stats NZ has a new website.

For new releases go to

www.stats.govt.nz

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Improving the household income and expenditure statistics

This document advises customers of the changes made to the Household Economic Survey (HES) 2015/16 and helps to understand the effect they may have when comparing the latest data with previous HES years.

The effect of these changes will be seen in Household Expenditure Statistics: Year ended June 2016, to be released 2 December 2016.

The need to make changes to HES

HES is an annual survey designed to measure the economic well-being of New Zealanders. While the survey measures household income and housing costs every year, we collect wealth and detailed housing expenditure only every three years. The current HES 2015/16 ran from July 2015 to June 2016 – to collect detailed household expenditure information and information on income and material well-being.

HES feeds into vital economic measures for the country, such as the consumers price index, household living-costs price indexes, and gross domestic product. HES data is also used to:

  • indicate the overall living standards and well-being of New Zealanders
  • advise on income inequality and poverty
  • help with decisions such as how changes to tax thresholds affect different types of households.

For these vital economic measures, it is important that HES data is accurate and reflects the real spending patterns and income of households.

We redesigned the 2015/16 HES questionnaire to take into account real-world changes, to address specific quality issues, and to improve respondents’ experience.

How these changes affect customers

While HES still collects the same information as in previous years, changes in how this information is collected may affect comparisons with previously published HES data. At a high level, these changes could mean:

  • expenditure levels for some expenditure groups will be different, due to survey changes. These included changes to: the period respondents report their expenditure for, the order the questions are asked, the question wording, and change in the split between which items are collected in the diary or questionnaire.
  • changes to housing costs to income ratios, due to the way we derive mortgage expenditure. We previously thought mortgage expenditure was under-estimated.

As a result of the changes made to the HES questionnaire, time-series analysis may not be possible in all expenditure and income categories. We had to balance time-series comparability with maintaining data integrity and reducing respondent burden. We have not been able to adjust previously published data because it has not been possible to unpick the many effects of the changes. The only exception is classification changes in income from self-employment, which we have revised back to 2007.

Changes made to HES 2015/16

For the 2015/16 survey we reworded and reorganised questions to make them easier for our respondents to understand and answer. We also made changes to address some known quality issues, such as the estimation of mortgage expenditure. Before being included in the final survey, changes to the questionnaire went through extensive cognitive testing and field testing.

In addition to the major changes in tables 1 and 2 below, we made minor changes across other HES survey questions. These included:

  • using plain English that is easy to understand for our respondents
  • reducing the level of detail asked of respondents, without compromising on data quality.

Table 1

Expenditure questionnaire (EQ): Changes made to questions relating to:

Household travel expenditure

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Respondents now asked to report expenditure on travel in the last 3 months. Reported expenditure is multiplied by 4 to get an estimate of travel expense over the 12-month period.
 Respondents were asked to report expenditure on travel in the last 12 months.  Remembering information over shorter periods is easier for respondents and enhances data quality.  A 3-month recall may increase response so may push up expenditure.

Household health expenditure

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Respondents now asked to report expenditure on health in the last 3 months. Reported expenditure is multiplied by 4 to get an estimate of expenditure over the 12-month period.
 Respondents were asked to report expenditure on health in the last 12 months.  Remembering information over shorter periods is easier for respondents and enhances data quality.  A 3-month recall may increase response so may push up expenditure.

Household maintenance

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Respondents now asked for total amount paid rather than itemised cost.  Respondents were asked for expenditure for each item used.  Recalling expenses by work done rather than by each item used reduces respondent burden.  There may be a shift in the components of maintenance expenditure.

Household mortgage expenditure

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Mortgage payments are now derived – based on current outstanding balance, interest rate, and last repayment amount.  Mortgage derivations were based on initial amount borrowed and did not allow for changes in the loan, such as lump sum payment or interest rate changes.   This change gives us a better estimate on mortgage expenditure.



 Previous mortgage figures were possibly under-estimated. This will have a bearing on housing-cost-to-income ratios.
 Mortgage questions reordered to be earlier in the expenditure questionnaire. Mortgage questions were asked towards the end of the expenditure questionnaire.
 When asked towards the end, mortgage questions often affected response quality. Moving it up will get a more-accurate estimate of mortgage expenses.
 

Collection of small regular items in diary

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Small regularly purchased items are captured only in the diary. Large irregular purchases are sourced from the EQ.
We have changed what is captured in the diary.
 Expenditure items were sourced from both the diary and the questionnaire and we had predetermined thresholds.  The thresholds were outdated.
Moving smaller items into the diary reduces respondent burden.
 Expenditure for items may increase or decrease depending on the source chosen.

Table 2

Income questionnaire (IQ): Changes made to questions relating to:

Contribution to saving schemes

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 Questions about contributions to saving schemes are asked in IQ.  These questions were previously asked in EQ.  Asking these questions with the income questions prompts better responses.
 Likely to result in higher reporting of contributions to saving schemes.

Self-employment

What is new?   What did we do in the previous HES?  Why make this change?  Likely impact on output data
 New structure enables respondents to report all the ways they get their income from their business. These could be profit or loss, wage or salary, and/or drawings.  Those reporting incomes from self-employment could report either ‘wage and salary’, or ‘profit and loss’. They could not report wage/salary with profit/loss from self-employment.  Total income for some self-employed respondents could have been under-reported.  Self-employment income is likely to be larger. However, the impact on total income is likely to be small. 



 Income of self-employed respondents, even if reporting salary/wages, is now classified as self-employment income.

 People who were self-employed but paid themselves a wage/salary were classed as wage and salary earners rather than self-employed.

 To improve the count of the number of people self-employed.  To account for the effect on the time-series comparison, we revised the data back to 2007.

Citation
Statistics New Zealand (2016). Improving the household income and expenditure statistics. Retrieved from www.stats.govt.nz.

ISBN: 978-0-908350-86-5

Page published 30 November 2016

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Top
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+