Stats NZ has a new website.

For new releases go to

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+
Household Economic Survey (HES) and Household Economic Survey (Income) resource

This page is a learning resource for the Household Economic Survey (HES). It outlines the purpose of the survey and its main uses, and provides a few practical exercises relating to the survey. 

What is the Household Economic Survey?

The Household Economic Survey (HES) provides a comprehensive range of statistics relating to income and expenditure, and personal and household demographics. The survey shows annual income from all sources at both the individual and household levels. An annual rather than a weekly measure of income tends to give a better indication of living standards, since an annual measure gives a longer-term view of income. The emphasis on households within the HES is also useful for assessing living standards because data includes the number of individuals a given income needs to support.

The HES collects information on each household's sources of income in great detail, and shows the percentage share each source contributes to household income. For example, in 2009/10 the HES showed that income from wages and salaries made up 74 percent of total household income.

The HES provides information about a significant number of demographic variables, or characteristics. For example, the HES shows how many people with specific characteristics, such as sex, ethnicity, labour force status, or age, live in households with a given range of household income.


How is the data collected?

HES data is collected from a sample survey. Sample surveys use a group of the population as representative of the whole population. HES data is collected from households – data on income is available for individuals or households, but expenditure can only be given for households.

The ‘full’ HES runs every three years. In this survey, data is collected through a household demographic questionnaire and a household expenditure questionnaire. Each eligible person is also given an income questionnaire and asked to complete an expenditure diary for two weeks. The household expenditure questionnaire collects information on regular payments such as mortgage, rent, and telephone, and also on big purchases, such as television sets. In the expenditure diary the respondent is asked to record everything they spent money on in the previous two weeks. The last full HES ran in 2009/10 and the next will run in 2012/13.

In the years between the full HES, starting in 2007/8, respondents are given the household demographic questionnaire, the income questionnaire, and a shortened household expenditure questionnaire (income) about certain basic housing costs such as rent, mortgage, and rates payments. This smaller survey is known as HES (Income).


How is the HES used?

Overall, the HES can be used to provide an indication of the overall living standard of New Zealand. The HES is useful for looking at the distribution of income in New Zealand. It also can be used to look at expenditure. Household expenditure is available for individual items or for groups of items.

Surveys collected from households typically provide rich demographic data such as information on ethnicity, qualifications, and household type. This means that household surveys enable comparison between particular demographic groups. At the same time, household information, such as the number of dependent children, offers a greater understanding of people's economic situation. For example, while an annual income of $60,000 may be considered a reasonably high income for a single person, it would not be so for a household with six dependent children.


What are the common confusions?

The HES collects data in great detail and, in order to reduce the burden on respondents it has a small sample size of approximately 4,700 households. This means that it is only possible to provide data at the national level, except for some large regions.

Due to the relatively small sample size, the HES can only report on what households spend, rather than providing a breakdown on individuals’ spending.

It is important to note that data is only comparable if it is taken from the same dataset or series. The HES collects data on both regular and irregular income. This means that HES income data cannot be compared with other surveys that only collect data on regular income (for example, the Labour Cost Index).


Statistical calculations


Sample – a subset of the population that is representative of the total population, and which therefore allows generalisations to be made about the total population (based on survey results).

The HES is a sample survey. Samples are designed to ensure that respondents represent people or households in the population that are similar to themselves, and are chosen scientifically but randomly. Random sampling increases the accuracy of the sample’s representation, because everyone has an equal chance of being selected. Larger samples also increase data accuracy. The HES has a sample of 4,700 households.

Other types of data collection are a census (where everyone in the population is asked questions on the topic) and outputs using administrative data (where data collected for other purposes is used for statistical purposes). Using school roll data to produce educational outputs is an example of how administrative data is used.


Example 1

To find the average weekly expenditure on bananas in their region, a student went to their local grocery store. They asked the first five people they saw buying bananas how much they spent on bananas that day. The student recorded how much the sample had spent on bananas on average and used this information to generalise across the whole region. Is this a representative sample? Why or why not?


No, it is not representative. It is likely that people shopping at the same grocery store live in the same area, which would mean they could have a very similar financial situation (for example, similar income). This means that the information cannot be generalised to people outside these characteristics. The student should also have counted people who didn’t purchase bananas in the sample. In addition, the results from such a small sample do not accurately represent the region’s population, as there is not enough randomness, or difference.

Example 2

Taking the scenario from example 1, what could be done to improve the sample design?


There are a number of ways in which this sample design could be improved, with a focus on two things – randomising the sample, and increasing the sample size. To randomise the sample, the student could go to different grocery stores around the region and ask every fifth person how much they spent on bananas. This would be more representative of the whole region, as using a variety of grocery stores would include more people with different demographic characteristics. To increase the sample size, the student could ask 10–20 people at each grocery store. With a sufficient number of stores, this would increase the sample size significantly.


Activity 1

A student wanted to know about the income and expenditure (spending) of New Zealand households for 2008. They decided the best way to find this information was to go to the public and ask for this data. The student went to two neighbouring households and asked everyone over 15 years of age about their income and expenditure amounts over 2008. There were six people in this sample. Is this sample representative of the whole New Zealand population? Why or why not?

Activity 2

Imagine you want to collect information on people’s opinions about how their spending habits (or expenditure) have changed since the beginning of the economic downturn. Consider the points made in the above summary on samples, then explain how you could choose an appropriate sample to collect this data.


Further reading

Household Economic Survey: Year ended 30 June 2010
User Guide for Statistics New Zealand’s Wage and Income Measures
Household Economic Survey (Income): Year ended 30 June 2009


Activity 1

No, it is not representative – for starters, this is too small a sample to use the results to make inferences about the whole population. More importantly, this is not a random sample. Two neighbouring households are likely to be very similar, especially with regard to labour force status, income, and expenditure. Therefore, the data from these two households will be very similar, which does not take into account differences between various groups within the population.

Activity 2

There are many possible ways to choose a sample. Answer should include:

  1. Randomly choosing people to be included in the sample.
  2. Having a large enough sample size to make sure there is enough, and sufficient variety of, information to represent the whole population.
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+