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+
Preview of 2016 national accounts improvements
Main updates to quarterly gross domestic product

This chapter outlines the improvements we made to the quarterly measures of economic growth.

Overview of quarterly GDP publication

GDP is New Zealand’s official measure of economic growth. The production measure of GDP (GDP-P) refers to the volume of goods and services produced in the economy, while the expenditure measure of GDP (GDP-E) shows how these goods and services were used. We express quarterly GDP-P in volume terms only (with the effects of price changes removed), and we express GDP-E in both volume terms and current price terms.

This chapter describes changes we have made to data and methodology that will affect our estimates of GDP growth. More detailed analysis of the changes and the revisions to total GDP growth will be published in the September 2016 quarter GDP release on 15 December 2016.

Annual benchmark updates

Updating annual benchmarks is essential for measuring quarterly economic growth. These updates ensure that our quarterly estimates capture changes in the structure of the economy. The quarterly growth rates are based on short-term indicators of change relating to a particular industry contribution to GDP-P, or an expenditure component of GDP-E. To measure overall growth, we weight the indicators of change according to the annual benchmark values derived in the national accounts. We update the weights used to measure economic growth as we update the national accounts, which includes the revisions outlined above.

Annual benchmark updates have a very significant effect over the long term. The impact of structural changes and updated data result in revised growth rates for the most recent periods. We are currently incorporating regular annual benchmarks for 2013 and 2014 into the GDP-P volume measure. The GDP-E statistics will also be updated with new 2013 and 2014 benchmarks, and additional benchmarks for 2015 and 2016. We will publish the effect of these changes in the September 2016 quarter GDP release on 15 December 2016.

Methodology improvements

In addition to the annual benchmark updates, we implemented improvements that led to revisions to GDP series. The largest of these is improvements in our methodology for measuring construction industry activity.

Construction industry improvements

The main change to the construction industry methodology is that we will benchmark the construction services sub-industry to annual value added (where previously there was no annual benchmark). The new annual volume benchmarks are created by deflating current price value added using the construction services output sub-index of the producer price index (PPI). This new methodology will cause revisions to annual construction volumes from the year ended March 1996, with annual growth estimates that show considerably less volatility on a year-to-year basis. Quarterly movements will be largely preserved.

The following figures (1a and 1b) indicate the changes to annual volume estimates that can be expected as a result of implementing this new method. The data in these figures does not yet include the effect of updated annual benchmarks. The updated benchmarks will result in additional revisions only for years after 2012.

Figure 1a

Graph, Effect of new construction services annual method on annual construction industry volumes, year ended March 1996 to 2016.  

Figure 1b

Graph, Effect of new construction services annual method on annual GDP volume growth rates, year ended March 1996 to 2016.

A number of smaller changes will also cause revisions to construction series:

  • Revisions to input data relating to heavy and civil construction in the March and June 2016 quarters will cause downward revisions to the construction industry, and investment in other construction.
  • Quarterly indicator data for residential building construction and residential building investment will be revised back to 2012, so that in future seasonally adjusted outputs will be more closely aligned to those published in the Value of Building Work Put in Place release.
  • The seasonal adjustment for the construction industry will change from indirect to direct.
  • Investment in other construction (including roading and telecommunications infrastructure) will now be seasonally adjusted, which will remove considerable volatility from the affected series.

Removal of fixed weights

We introduced chain-linking to replace fixed 2010 weights for lower-level series in the mining and business services on the production side of GDP. This means that we now annually update the current price weights used to aggregate sub-industries into industry totals (see Quarterly gross domestic product: Sources and methods (fourth edition) for further detail on chain linking). We also introduced chain-linking of FISIM into the export and import components, and some parts of non-residential gross fixed capital formation, on the expenditure side of GDP.

Central government final consumption expenditure

Revisions to central government final consumption expenditure are primarily due to updates to data and classifications, along with removing the small number of remaining fixed 2010 weights. The changes result in revisions to the quarterly movements of both current price and volume estimates from June 1994 onwards. They change the GDP-E quarterly seasonally adjusted movements by up to 0.1 percentage points. They do not affect the annual movements as the annual benchmarks are unchanged. There is no impact on GDP-P.

Changes to data sources

We made several small changes to data used for quarterly GDP:

  • We aligned oil and gas production indicators more closely with those published in MBIE’s Quarterly Energy Report.
  • Updated travel credit proportions have been incorporated into our measures of household consumption expenditure, although these changes have no effect on household consumption totals.
  • We updated the annual data for school education (for the period from 2011 to 2015), with data sourced from the Ministry of Education. This replaces the previous estimates, which were derived from a population-based method.

All changes to quarterly GDP methodology will be reflected in the fifth edition of the quarterly GDP sources and methods, due to be released at the end of 2016.

  • 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+