³Ô¹ÏÍøÕ¾

This page provides information on changes in the methodology used to calculate seasonal adjustment for series in business financial data (BFD) and the component surveys associated with it, that is, the Economic Survey of Manufacturing (MFG), the Wholesale Trade Survey (WTS), and the Selected Services Survey (SSS). The Retail Trade Survey (RTS) has already been through an update of its seasonal adjustment methodology. However, there are two series derived from RTS data that are only produced as part of the BFD, which are addressed in these updates.

These updates are being made to better account for the impact of COVID-19, and to address changes in the seasonality of each series since the redesign of the surveys in business financial data released in June 2016. These enhancements will take effect from the publication of Business financial data: September 2024 quarter on 11 December 2024.

On this page

• Overview
• Background to seasonal adjustment
• Additive outliers for seasonal adjustment and trend
• Changes to treatment of aggregated series
• Changes to series with/without seasonal patterns
• Changes to series with temporary disruption in seasonality
• Changes to series affected by structural changes
• Comparison with GDP
• Impact on the data

Overview

We are implementing multiple enhancements to the seasonal adjustment methodology for series related to quarterly business financial data and its component surveys.

We used these updates to:

• improve accuracy due to more appropriate treatment of COVID-19-affected periods
• improve our approach to extreme outliers using the new Stats NZ standards for additive outliers in seasonal adjustment
• improve alignment of treatment with related Stats NZ outputs, notably the Retail Trade Survey (RTS) and gross domestic product (GDP)
• resume publishing trend series, which had been suppressed during the periods significantly affected by COVID-19-related impacts
• identify series or certain periods of series for which seasonal adjustment is no longer appropriate
• evaluate whether seasonal adjustment is now appropriate for series which were previously not adjusted
• identify aggregated series for which indirect seasonal adjustment is appropriate
• identify other treatments such as applying level shifts, to account for real world changes affecting any series.

These enhancements will cause revisions to the seasonally adjusted and trend series back to June 2016 quarter (inclusive). While seasonally adjusted data is revised with each new quarter of data, these updates will cause larger than normal revisions. Details on the impact of these changes and the reasons behind them are outlined in the following sections.

We have included five files to help understand them:

• CSV files with the data before and after implementation of the new methodology up to the June 2024 quarter
• an Excel workbook listing seasonal adjustment treatments before and after implementation.

See Download data for the relevant files.

The following example highlights the impact of the change.

On 10 September 2024 we published a 0.6 percent change in the volume of total manufacturing sales for the June 2024 quarter compared with the March 2024 quarter. Using the updated seasonal adjustment methodology, the June quarter volume of total manufacturing sales would have shown a rise of 0.3 percent.

This new methodology will be used from our next publication Business financial data: September 2024 quarter on 11 December 2024.

This methodology change has no impact on our actual data (not seasonally adjusted), on our deflators, nor on any other Stats NZ collection. While our data is used by other collections (for example, in quarterly GDP estimation), we independently calculate seasonal adjustment across collections using comparable methodology. This maintains consistency where possible while recognising the unique nature of each series.

Background to seasonal adjustment

The majority of series within BFD and its component surveys exhibit seasonal patterns. Stats NZ performs seasonal adjustment on our time series. Seasonal adjustment in Stats NZ has more information on how we estimate and remove the seasonal effects from the time series.

Since 2020, we have applied an additive outlier treatment to series within BFD and its component surveys where there were major impacts from border closures or domestic COVID-19 restrictions. This subdued the impact of unusual data points on the seasonal adjustment process.

The impact of COVID-19 on the BFD series has now largely diminished for the most recent quarters. Therefore, we are updating our approach to seasonal adjustment of COVID-19-affected periods and introducing the new Stats NZ standard for treating extreme outliers using additive outlier treatment.

Seasonal adjustment and automatic outliers in time series after COVID-19 has more information on the new Stats NZ approach to additive outliers.

These changes will also improve our alignment with other Stats NZ releases like the RTS released on 16 February 2024 and GDP (Improvements to gross domestic product September 2024 quarter) to be released on 27 November 2024.

While improving our approach to extreme outliers and reintroducing trends, we took the opportunity to review other aspects of our seasonal adjustment methodology. We last redesigned BFD and its component surveys in June 2016, and we generally need a five-year time series to assess seasonality, so now was a good time to look at a variety of enhancements to our seasonal adjustment methodology.

Additive outliers for seasonal adjustment and trend

Seasonal adjustment and automatic outliers in time series after COVID-19 describes how we previously managed extreme outliers using an additive outlier method at Stats NZ, and the methodology changes we are now implementing for series within BFD and its component surveys

We have applied methodology consistent with other Stats NZ collections wherever possible. However, series within BFD and its component surveys have been thoroughly analysed to ensure the most appropriate treatment to be used for each unique series. We have continued to make use of manual additive outliers during the most significantly affected COVID-19 period (March 2020 to December 2022 quarters, inclusive) and implemented automatic detection from the March 2023 quarter onwards. Automatic outlier detection can occasionally change which quarters it assigns as an outlier when data from a new quarter is added to a time series. Manually assigning outliers during the COVID-19 period ensures the selected outliers will not change as more data is added. This reduces the magnitude of revisions to seasonally adjusted data. We have aligned our treatment of additive outliers with quarterly GDP to ensure consistency across our collections.

Revision of the manual adjustments and use of some other complex treatments to manage impacts through COVID-19-affected periods, mean we have improved the quality of trend series to the point they can be published again, as from the March 2020 quarter.

A list of outliers applied to each series before and after our updates is available in the Excel file ‘Additive outliers and complex treatments, old and new methodology’ under Download data.

Several other Stats NZ collections have already adopted our new internationally aligned standards for additive outlier detection, including the consumers price index, Household Labour Force Survey, Quarterly Employment Survey, value of building work put in place, and RTS. All our collections will eventually align with the new standards, after we have analysed the methodology in the context of each unique series.

Changes to treatment of aggregated series

We seasonally adjust aggregate series using a mixture of direct and indirect seasonal adjustment. Direct seasonal adjustment involves seasonally adjusting the direct sum of an aggregate’s component series, while indirect seasonal adjustment involves summing the seasonally adjusted estimates of the relevant component series. The indirect method is generally preferred as there is no discrepancy between the seasonally adjusted values for aggregated series and the sum totals for its components.

We analysed the aggregated series for direct and indirect seasonal adjustment and found that there was no significant difference between the two measures. We chose to implement indirect seasonal adjustment to meet our customer preferences and to maintain consistency in our approach with GDP where appropriate.

Table 1 shows the series that have switched to indirect seasonal adjustment.

Table 1, Series changed from direct to indirect seasonal adjustment, by industry and series reference

Industry	Series reference	Seasonally adjusted values/volumes	Variable	Previous method	New method
Agriculture, forestry, and fishing	BDCQ.SF1AACS	current values	sales	direct	indirect
All manufacturing	MFGQ.SFZ1CS*	current values	sales	direct	indirect
All manufacturing	MFGQ.SFZ2CS	current values	purchases	direct	indirect
All manufacturing	MFGQ.SFZ2KS	volumes	purchases	direct	indirect
All manufacturing excluding meat and dairy	MFGQ.SFY1CS	current values	sales	direct	indirect
All manufacturing excluding meat and dairy	MFGQ.SFY1KS	volumes	sales	direct	indirect
All manufacturing excluding meat and dairy	MFGQ.SFY2KS	volumes	purchases	direct	indirect
Food, beverage, and tobacco product manufacturing	BDCQ.SF1CC1CS	current values	sales	direct	indirect
Meat and dairy product manufacturing	MFGQ.SFA1CS	current values	sales	direct	indirect
Meat and dairy product manufacturing	MFGQ.SFA2CS	current values	purchases	direct	indirect
Meat and dairy product manufacturing	MFGQ.SFA1KS	volumes	sales	direct	indirect
Meat and dairy product manufacturing	MFGQ.SFA2KS	volumes	purchases	direct	indirect
Total wholesaling	WTSQ.SFZ1CS**	current values	sales	direct	indirect
Professional, scientific, technical, administrative and support services	BDCQ.SF1MNCS	current values	sales	direct	indirect
Arts, recreation and other services	BDCQ.SF1RSCS	current values	sales	direct	indirect
* BDCQ.SF1CCCS is populated from this series and values are equal. ** BDCQ.SF1FFCS is populated from this series and values are equal.

Aggregate series which switched to indirect seasonal adjustment are available in the Excel file ‘Additive outliers and complex treatments, old and new methodology’ under Download data.

Changes to series with/without seasonal patterns

It is best practice to regularly review seasonal adjustment methodology, and it generally requires a minimum of five years’ data to identify new patterns. Introducing changes to additive outliers due to COVID-19 impacts gave us the opportunity to look at seasonality changes that have emerged since our June 2016 redesign.

We have identified four series in the MFG and two series in the WTS that no longer show a significant seasonal pattern, and we have therefore stopped seasonally adjusting these. The data will still be published under these seasonally adjusted series in Infoshare, but it will be identical to the unadjusted data starting from the June 2016 quarter. Trend series will still be produced and published.

Table 2 shows the series that are no longer seasonally adjusted.

Table 2, Series no longer seasonally adjusted, by industry and series reference

Industry	Series reference	Seasonally adjusted values/volumes	Variable	Previous method	New method
Seafood processing	MFGQ.SFB1CS	current values	sales	seasonally adjusted	not seasonally adjusted
Seafood processing	MFGQ.SFB1KS	volumes	sales	seasonally adjusted	not seasonally adjusted
Seafood processing	MFGQ.SFB2CS	current values	purchases	seasonally adjusted	not seasonally adjusted
Seafood processing	MFGQ.SFB2KS	volumes	purchases	seasonally adjusted	not seasonally adjusted
Total wholesaling	WTSQ.SFZ1CS	current values	stocks	seasonally adjusted	not seasonally adjusted
Motor vehicle and motor vehicle parts wholesaling	WTSQ.SFC9CS	current values	stocks	seasonally adjusted	not seasonally adjusted

We have identified two series in BFD and three series in MFG that show evidence of seasonality, and we have begun seasonally adjusting them. Table 3 shows the series that are now seasonally adjusted.

Table 3, Series now seasonally adjusted, by industry and series reference

Industry	Series reference	Seasonally adjusted values/volumes	Variable	Previous method	New method
Fruit, oil, cereal, and other food manufacturing	MFGQ.SFC2CS	current values	purchases	not seasonally adjusted	seasonally adjusted
Fruit, oil, cereal, and other food manufacturing	MFGQ.SFC2KS	volumes	purchases	not seasonally adjusted	seasonally adjusted
Non-metallic mineral product manufacturing	MFGQ.SFJ2CS	current values	purchases	not seasonally adjusted	seasonally adjusted
Arts, recreation and other services	BDCQ.SF1RSCS	current values	sales	not seasonally adjusted	seasonally adjusted
Other services	BDCQ.SF1RS2CS	current values	sales	not seasonally adjusted	seasonally adjusted

The following example graph depicts seasonal patterns for Fruit, oil, cereal, and other food manufacturing.

A list of series with these changes is available in the Excel file ‘Additive outliers and complex treatments, old and new methodology’ from Download data.

Changes to series with temporary disruption in seasonality

From early 2020, the COVID-19 pandemic saw governments around the world, including New Zealand, impose international travel restrictions due to the spread of COVID-19. New Zealand’s border restrictions were gradually eased from late February 2022, and the border fully reopened on 1 August 2022. Social distancing and border closures severely impacted four series in BFD and disrupted their regular seasonal patterns. We have decided to not seasonally adjust these series through the affected period. Data will still be published under the seasonally adjusted series for the affected period, but it will be identical to the actual data. To ensure a smooth transition back to seasonal adjustment, we have created a synthetic series for the affected period. This synthetic series will not be published but will be used to determine the seasonal factors.

Methodology for bridging seasonally adjusted international travel and migration series impacted by COVID-19 has more information on the Stats NZ approach to bridges for seasonal adjustment.

Table 4, Bridge period for four BFD series

Industry	Series reference	Bridge period (inclusive of quarters)
Accommodation and food services	BDCQ.SF1GH2CS	June 2020 quarter to September 2022 quarter
Transport, postal and warehousing	BDCQ.SF1IICS	March 2020 quarter to June 2022 quarter
Administrative and support services	BDCQ.SF1MN2CS	March 2020 quarter to September 2022 quarter
Arts and recreation services	BDCQ.SF1RS1CS	March 2020 quarter to June 2022 quarter

Changes to series affected by structural changes

While reviewing our automated outlier processes during the most COVID-19-affected periods, we identified some structural changes in specific industries which could be better managed with the introduction of complex treatments: level shifts or temporary changes. These structural changes are not necessarily related to COVID-19 disruptions but happened during the same period. As other structural changes are observed in the future, more level shifts or temporary changes may be applied.

Seasonal adjustment and automatic outliers in time series after COVID-19 has more information on level shifts and temporary changes.

The closure of Marsden Point Refinery saw large oil companies change their Australian and New Zealand Standard Industrial Classification (ANZSIC) from CC511 (petroleum and coal product manufacturing) to FF111 (petroleum product wholesaling). This caused a large level shift for both industries. The series in table 5 have level shifts applied to them in June 2022.

Table 5, Series with level shifts applied in June 2022, by industry and series reference

Industry	Series reference	Seasonally adjusted values/volumes	Variable	Level shift
Petroleum, chemical, polymer and rubber product manufacturing	BDCQ.SF1CC5CS *	current values	sales	June 2022
Petroleum and coal product manufacturing	MFGQ.SFH1CS	current values	sales	June 2022
Petroleum and coal product manufacturing	MFGQ.SFH1KS	volumes	sales	June 2022
Petroleum and coal product manufacturing	MFGQ.SFH2CS	current values	purchases	June 2022
Petroleum and coal product manufacturing	MFGQ.SFH2KS	volumes	purchases	June 2022
Total wholesaling	WTSQ.SFZ1CS *	current values	sales	June 2022
Total wholesaling	WTSQ.SFZ9CS	current values	stocks	June 2022
Basic material wholesaling	WTSQ.SFA1CS *	current values	sales	June 2022
Basic material wholesaling	WTSQ.SFA9CS	current values	stocks	June 2022
* Level shifts had been implemented previously for these series and have been reviewed.

Comparison with GDP

As data collected under the BFD and its component surveys is used as a major component of the quarterly GDP measure, we have worked closely to align our updates to that of quarterly GDP where appropriate.
Overview of sources and methods for quarterly gross domestic product: Updates and COVID-19 adjustments has more information on the data sources used in the calculation of quarterly GDP.

Impact on the data

The collective impact of all these enhancements will cause revisions to the seasonally adjusted and trend series as far back as the June 2016 quarter (inclusive). This was when BFD and its component surveys was last redesigned.

The biggest differences are observed in the COVID-19-affected periods (March 2020 to September 2022), and in the most recent periods (2024 quarters) as the latest data points in a series are usually most prone to revision.

CSV files for all surveys are available under Download data for comparison. These contain the complete dataset up to the June 2024 quarter, both under the previous and the new methodology. These will allow you to see the impact of the methodology change, without the additional revisions to seasonally adjusted data that are brought in by a new quarter’s data. These figures will be revised, as is normal seasonal adjustment practice, with publication of the September 2024 quarter results on 11 December 2024.

ISBN 978-1-991307-37-8

Enquiries

Divya Sharma
04 931 4600
[email protected]