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 living-costs price indexes – decisions after public consultation
Decisions we've made


Table 1 outlines the decisions we made for specific questions we consulted about.

Table 1 Summary of decisions about HLPIs 
Question  Decision
2.1 Beneficiary household definition Highest-income recipient receives a ‘main benefit’
2.2 Superannuitant household definition  Highest-income recipient receives a NZ government pension 
2.3 Māori household definition At least one person reports Māori ethnicity 
2.4 Quintile definitions 

Equivalised disposable income, using modified OECD scale

Additionally, as separate indexes

Equivalised expenditure  

3.1 Aggregation method  Democratic weighting (to reflect expenditure patterns for a ‘typical’ household in each group)
4.1 Interest payments method

All consumer debt in scope

Property price index as debt index

Simple revaluation approach  

4.2 Expenditure on items insured  Gross expenditure on items insured 
4.3 Second-hand goods expenditure  Net expenditure on second-hand goods 

Defining household groups

The eight household groups chosen for HLPIs were the ones most commonly identified in submissions to our public consultation (Statistics NZ, 2014). Classifying households into HLPI groups will be based on demographic information we collect in the Household Economic Survey (HES).

See Household Economic Survey for more about what the survey covers.

Table 2 summarises the household definitions and outlines some alternatives we considered.

Table 2 Household group definitions
Household group Current working definition  Potential alternatives 
Beneficiary Households where the highest-income recipient receives a benefit payment, classified as a ‘main benefit’ in HES.1  

Benefits classified as ‘supplement benefit’ in HES could also be included.

Households where a minimum proportion of income (eg 75 percent), comes from benefits.  

Superannuitant  Households where one or more people aged 65+ receive a New Zealand government pension, and where a superannuitant had the highest household income. 

Households where the highest-income recipient received a New Zealand government pension (no age criteria).

Households where a minimum proportion of income (eg 75 percent), comes from a New Zealand government pension. 

Māori Households where at least one member has reported Māori ethnicity (as one of their ethnicities).  Using Māori descent would be possible in future, as a descent question was added to HES from the 2015/16 survey.  
Income groups  Households classified into five income groups (quintiles), using before-tax regular and recurring income. 

Possibilities include:

  • equivalised income
  • disposable income
  • expenditure quintiles.  
1 This group definition was referred to as ‘main beneficiary’ in the 2013 feasibility study.

For each household group we asked customers what definition they thought was most appropriate and why.

Additionally, for the income groups we asked customers whether:

  • they thought income equivalisation should be used and why
  • gross or disposable income should be used
  • we should use expenditure rather than income to define the groups.

Beneficiary group definition

The definition for a beneficiary household that we’ll use is ‘the highest-income recipient in a household receives a ‘main’ benefit’. All submissions preferred this definition.

Superannuitant group definition

The definition for a superannuitant household that we’ll use is ‘the highest-income recipient in a household receives a New Zealand government pension’. This was the most-common preference and is a simple definition that we will not need to revise if the pension age changes.

We also asked what pensions should be classified as New Zealand government pensions. We decided to stay with the working definition, which includes the following:

  • New Zealand Superannuation paid by Work and Income
  • Veteran’s pension
  • war disablement pension
  • surviving spouse pension
  • overseas pension top-up payment from Work and Income
  • overseas pension paid by Work and Income using the special banking option
  • other type of New Zealand government pension.

We saw this as the simplest approach and it is also most consistent with HES.

Māori group definition

We will use a wide definition for Māori household – it will be one where any household member reports Māori ethnicity as one of their ethnicities. We received support for this definition and it is consistent with the one used in recent publications by Te Puni Kōkiri (Te Puni Kōkiri, 2011).

Following the consultation, we did some work to see how many households would be defined as Māori if we used a different definition of household ethnicity.

We explored the following three definitions of Māori households.

  • Anyone in the household reported Māori ethnicity (chosen definition).
  • Majority of people in the household reported Māori ethnicity.
  • Everyone in the household reported Māori ethnicity.
Table 3 Alternative definitions of Māori households
Definition  Proportion of households
Average household size
Median equivalised disposable income
(modified OECD scale)
Anyone Māori 17.4  3.1  27,200 
Majority Māori  15.2  3.0  26,500 
Everyone Māori  8.6  2.9  22,800 

Table 3 shows, using the 2009/10 HES, that a similar proportion of households are classified as Māori using the first two definitions. We also found similar expenditure patterns (and therefore price change) between these two groups. Average household size, and age (not shown), are similar for all three alternatives.

The third definition, where everyone in the household reported Māori ethnicity, results in the lowest median income. Over the study period 2008–12, the third definition results in the highest annual average price change (of the three alternatives). This result is consistent with the finding that lower income households had greater inflation over this period (see Bentley, 2015).

The relatively small proportion of households defined by the third definition means that the HES sample size is about half that of the alternatives we considered (it resulted in fewer than 200 households in recent HES). If the third definition was used, the lower sample size would make estimates of the group expenditure weights less reliable.

We will assess the sensitivity of the definition to using ethnicity, rather than descent, once Māori descent information is available from the 2015/16 HES.

Income group definition

We asked customers how they would use income group price indexes. The most common suggested use is to compare the inflation experience of the different groups. In particular, to understand how inflation for the lower two groups compares with the higher-income groups.

We will use income equivalisation – all submissions agreed with this approach. We’ll use the modified OECD scale to do this as it is simple, consistent with international best practice, and was the most common preference from the submissions.

We’ll use disposable (or net) income – that is, income available after income tax. This was the most popular choice as it represents what is available to spend on consumption.

Most customers were interested in income as a proxy for understanding standard-of-living, but also noted consistency with other income studies as being important. We received diverse views on the relative importance of income, expenditure, or joint income and expenditure quintiles.

We’ll also publish five additional indexes using expenditure rather than income quintiles. This is in response to several submissions suggesting a need for both income and expenditure group indexes. 

Aggregation method

We asked customers whether they thought it was more important to understand the inflation experience of households overall or of a ‘typical’ household within each group.

The majority of submissions specified they are most interested in a ‘typical’ household. Based on this, we’ve decided to use democratic weighting when calculating the expenditure weights for the HLPIs.

The weighting methods we use for the CPI involve calculating expenditure patterns from aggregate household expenditure – that is, calculating total expenditure on each commodity in the CPI basket for the CPI reference population (all private New Zealand-resident households). Effectively, this treats the whole of New Zealand as a single ‘super-household’. This is known as plutocratic weighting.

Plutocratic weighting is conceptually appropriate to the CPI’s principal use as a macroeconomic indicator for monetary policy purposes. However, it may be more appropriate for HLPIs to focus on the inflation for a ‘typical’ household within household groups. We can estimate inflation of a ‘typical’ household within a group by calculating expenditure proportions for each household and then taking a simple unweighted average of these proportions. This is known as democratic weighting.

Practical application of a ‘payment’ approach

The HLPIs will be constructed using a ‘payment’ conceptual approach. The International Labour Organization have noted that this approach is often used “when the primary purpose of the index is for the adjustment of compensation or income” (ILO, 2003). The 2013 CPI Advisory Committee recommended we use this approach for HLPIs.

The payment approach tracks the price change for goods and services paid for, regardless of when they’re acquired or used. Note: this does not include non-monetary consumption (eg food or fishing for self) or benefits in kind (paid to service providers).

In contrast, the CPI uses an ‘acquisition’ approach, reflecting price changes for goods and services when they are acquired.

The main practical difference between the two approaches is on measuring housing, interest, and insurance. The acquisition-based CPI commodity-level price indicators can be translated to a payment approach with the following modifications:

  • including interest payments
  • excluding net acquisition of owner-occupied housing
  • using gross expenditure weights for insurance.

We asked customers for their views on:

  • what type of consumer debt should be included in the HLPIs
  • what debt index should be used to quality adjust nominal interest rates
  • what debt revaluation approach we should use
  • whether a gross approach to insurance is appropriate
  • whether a net approach to second-hand goods is appropriate.

Interest payments

We’ll use all consumer debt for the HLPIs. This includes mortgage debt for housing costs, mortgage debt for other purchases, and other forms of consumer debt (eg credit cards, hire purchase, and car loans). All the submissions agreed this is the most appropriate scope of debt to include.

We’ll use a property price index (Property IQ’s house price index) to quality adjust the nominal interest rate of housing debt and the CPI to quality adjust all other debt. Submissions that supplied a response to this question supported this view.

We’ll use a simple revaluation approach – the most common view of the submissions.


For HLPIs, under the ‘payment’ approach insurance expenditure weights will be on a ‘gross’ basis, based on total household spending on insurance premiums. This is in contrast to the ‘net’ basis used in the CPI, which includes only the proportion of insurance premiums that contributes to the cost of providing the insurance service (ie premiums less claims).

This means the conceptual scope for insurance payments is the same as that used in HES; there will be no need for conceptual adjustments to determine the HLPI weights on insurance. However, we needed to determine the appropriate conceptual treatment of expenditure on items insured. This includes how to treat expenditure by insurance companies, on behalf of households, on replacement goods.

We received mixed views on this question – reflecting the dilemma the options posed. We’ve decided to use the gross expenditure on items insured, which is considered “the most appealing approach” by the Consumer Price Index Manual (ILO, 2004).

Second-hand goods

We’ll use net expenditure on second-hand goods. Most submissions preferred this approach.

The issue with second-hand goods is that these purchases don’t represent ‘final consumption expenditure’. In the CPI, second-hand purchases from businesses are in scope, as they represent new consumption for the household sector. And if purchases are from another household, expenditure is already captured – it could be considered double-counting to re-record expenditure on these products.

HES, used to derive expenditure weights, captures households’ expenditure on second-hand goods. The survey also captures the proceeds received by households from selling second-hand goods. This makes it straightforward to estimate households’ net expenditure on second-hand goods.

For the HLPIs, applying the payment approach will capture the net expenditure of the household group on second-hand goods – on goods that are ‘new’ to the group. The Consumer Price Index Manual (ILO, 2004) endorses this approach. 

Other outcomes

We received feedback about two further issues alongside responses to the questions we asked in the public consultation.

The first suggested that we add the costs associated with retirement villages to the basket of goods and services. This is particularly relevant for the superannuitant index. In response to this we’ll assess this possible basket addition as part of the next CPI review (due to be published with the September 2017 quarter CPI release in October 2017).

The second point was that it would be useful to publish an all-households HLPI so customers can compare the sub-populations and all New Zealand households. The benefits need to be balanced against concerns that an alternative ‘all-households CPI’ could be confused with the CPI.

We’ll continue to monitor interest in an all-households HLPI and welcome further feedback on this issue. 

  • 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+