|Page tools: Print Page Print All
CHANGES TO PRIVATE NEW CAPITAL EXPENDITURE AND EXPECTED EXPENDITURE STATISTICS
This information paper outlines changes in the frame definition and sample design for the Survey of New Capital Expenditure, taking effect in survey outputs from the release of the September quarter 2009 publication Private New Capital Expenditure and Expected Expenditure (cat.no.5625.0).
An important change is the implementation in the Survey of the Australian and New Zealand Standard Industrial Classification (ANZSIC), 2006, replacing the 1993 version of ANZSIC. The implementation of the ANZSIC 2006 industry classification framework is taking place in a coordinated program across ABS economic collections.
This paper is complementary to more general information papers ANZSIC 2006 Implementation, 2006 (cat.no.1295.0) and Update on ANZSIC 2006 Implementation (cat.no.1295.0.55.001) which outlined plans for the integration of ANZSIC 2006 across a range of statistics produced by the ABS.
Further information about the Survey of New Capital Expenditure can be accessed in the Explanatory Notes of Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0).
For information concerning changes to the Survey of New Capital Expenditure, described in this Information Paper, please call Paul Doran on (02) 9268 4357 or email <email@example.com>.
For more information on the ANZSIC 2006 and its implementation across a range of ABS statistics, please contact Celia Quiatchon, ANZSIC Implementation Manager, on telephone (02) 6252 5604 or email <firstname.lastname@example.org>.
This information paper outlines changes in the frame definition and sample design parameters for the Survey of New Capital Expenditure, taking effect in survey outputs from the release of the September quarter 2009 publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0).
These changes include:
These changes will have statistical impacts on Survey of New Capital Expenditure estimates. These impacts will be managed for users by the production and release of revised historic series which align past published estimates to the estimates for September quarter 2009 and onwards. This Information Paper describes more specifically (see Section 4) the approach taken to measure and integrate these statistical impacts into historic data series.
A sample copy of the table affected by these changes which will appear in the September quarter 2009 publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) is included in the PDF version of this information paper along with a copy of the previous version of this table. The template for each Time Series Spreadsheet is available under the 'Downloads' tab for this publication, at www.abs.gov.au. These sample tables show how the capital expenditure series will be published on an ANZSIC 2006 industry basis.
June quarter 2009 data are the final capital expenditure statistics available on the basis of the previous frame definition and sample design. There will be no adjustment of capital expenditure data of future periods to the previous design parameters.
A further change will be introduced for the September quarter publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0). The practice of publishing five year average realisation ratios for capital expenditure actual and expected statistics will be discontinued. However the complete time series of past realisation ratios are published on the web page for Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) at www.abs.gov.au as tables 12A and 12B. The experimental projection graphs and commentary, which are modelled on the basis of five year average realisation ratios will also be discontinued. For further information see Section 6.
2 CHANGES TO INDUSTRY CLASSIFICATION
The background to the development and implementation of Australian and New Zealand Standard Industrial Classification, (ANZSIC), 2006 (cat. no. 1292.0) is discussed in detail in the Information Paper Update on ANZSIC 2006 Implementation (cat. no. 1295.0.55.001). ANZSIC 2006 improves the currency and relevance of the classification of ABS data by industry and maximises the global comparability of Australian industry statistics by better alignment of the industrial classification framework with international standards.
The program for implementing ANZSIC 2006 into ABS industry statistics was developed to facilitate the release of the Australian National Accounts (ANA) on the new basis from late 2009. The change for the Survey of New Capital Expenditure to ANZSIC 2006 is part of a coordinated program of changes for statistical collections which supply data to the National Accounts.
First release of new outputs
The first dissemination of Survey of New Capital Expenditure outputs on an ANZSIC 2006 basis will occur from the September quarter 2009 data release on 26 November 2009, through the publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) and related time series data files which are available to users through the ABS web site www.abs.gov.au.
Changes to Industry classification of outputs
The key changes for detailed industry outputs from the Survey of New Capital Expenditure associated with the implementation of ANZSIC 2006 affect Table 2 of Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) and the Time Series Spreadsheets 2a - 2f available at www.abs.gov.au. The structure of Table 2 in the publication can be reviewed in the PDF version of this information paper.
The previous detailed industry publication framework was created in a period of lower private industry contribution in parts of Other Services. These parts of Other Services have subsequently increased in significance for private capital expenditure statistics following both privatisation and growth in these industries. The scope of the Survey has not changed in terms of including private and excluding government businesses.
In summary, the number of industries included in the publication of detailed industries for capital expenditure data has been expanded from nine to twelve and more closely follows the ANZSIC divisional structure. The changes are described in more detail in Section 6.
Minor scope changes have also occurred, with the inclusion of Waste Services industries as part of ANZSIC 2006 Division D and the exclusion of ANZSIC Division O Public Administration and Safety, which is predominantly non private enterprises, though historically was included in the Survey for private businesses which contribute to the Division.
This shift in the approach to publication of detailed industry capital expenditure survey outputs, consistent with ANZSIC 2006, has required different parameters in sample design for the Survey and is a key change for the Survey in respect to frame definition and sample design.
3 ADDITIONAL CHANGES
As noted in the preface to this paper, other changes for the Survey, including improving the Survey’s frame by the addition of non-employing businesses, the integration of updated size variable information for each unit and the complete enumeration of economically significant Unincorporated Joint Ventures are being made at the same time as the introduction of ANZSIC 2006. These changes are discussed in the following sections.
Inclusion of Non-Employing Businesses
The inclusion of non-employing businesses which contribute significantly to economic activity is a change in the design of the Survey aimed towards more complete coverage of the economy with respect to measurement of capital expenditure.
The non-employing businesses added to the frame for the Survey have been included according to the significance of their turnover relative to the industry in which they operate. Specifically, these businesses were measured by their turnover for Business Activity Statement (BAS) purposes, which was compared to total BAS turnover for the industry.
There remains a substantial number of business units operating under an Australian Business Number which are considered ‘micro non-employers’ and are not included in the Survey as it is considered that their contribution to capital expenditure would be insignificant.
Updated Size Variable Information for Businesses on the Survey frame
Information pertaining to the size of each unit on the ABS frame of businesses that is used across economic collections is an important design parameter for stratification purposes. A measure for employment is used for ABS economic surveys in this respect. There are weaknesses to using a count of a business’ employees so a Derived Size Benchmark is calculated from information available to ABS through the Australian Taxation Office Business Activity Statement data.
As part of the suite of design changes being implemented from September quarter 2009, the Derived Size Benchmark (DSB) for each unit on the ABS frame of businesses has been updated according to newly available information and a refined methodology. This has improved the Survey frame by making the categorisation of businesses by their size more representative of their economic contribution. In turn, this leads to more effective representation of selected units for others in each sample stratum. This new size variable is termed Stratification Derived Employment Size (SDES).
Complete enumeration of Unincorporated Joint Ventures
A further change to the frame definition and sample design for the Survey is the complete enumeration (that is, selection of these units with certainty and with a sample weight of 1) of Unincorporated Joint Ventures (UJVs) which contribute significantly to economic activity.
Previously, not all UJVs on the survey frame were selected in each quarter of the Survey. The decision to completely enumerate significant UJVs minimises an area of risk in terms of the sampling error on survey estimates.
Changes under Standard Economic Sector Classification (SESCA) 2002
The Standard Economic Sector Classifications of Australia (SESCA) (cat no. 1218.0) are a suite of classifications which are applied to institutional units. The main driver for the review of SESCA was the implementation of System of National Accounts 2008 (SNA08). The first SESCA classification reviewed was the Standard Institutional Sector Classification of Australia (SISCA).
The following changes were made to SISCA following the review:
The ABS made a commitment to users to implement the changes stemming from the introduction of internationally agreed enhancements to National Accounts Standards (SNA08) to coincide with the introduction of the new industry classification, the Australian and New Zealand Standard Industrial Classification 2006 (ANZSIC 2006), in order to avoid having two significant changes to statistical outputs within a short period of time.
Changes to the Standard Economic Sector Classification of survey units, under SESCA 2002 (cat. no. 1218.0) will have a minimal impact on capital expenditure estimates and are described for completeness in the description of all changes impacting the Survey.
4 IMPACT ON TIME SERIES
The changes described in this paper, being introduced for the Survey of New Capital Expenditure commencing September quarter 2009, will have statistical impacts on survey estimates. The comparability to published time series of both actual and expected capital expenditure are affected.
These impacts are being managed for users by the production and release of revised historic series which align past published estimates to the estimates including these statistical changes. Published series, including tables available on www.abs.gov.au are being revised from the inception of each series.
Methodology for backcasting survey estimates
Data has been collected from a parallel sample of units on the new frame and sample design basis through March and June quarters 2009. The parallel sample was a separate but overlapping survey sample. Parallel estimates have been compiled and analysed but not published through March and June quarters 2009. This has enabled the measurement of statistical impacts. These data will be used to backcast the new estimates to create revised historic series.
The ABS constructed historical ANZSIC 2006 base series for the quarters from September 2006 to December 2008 at the state by industry division level and for Division C Manufacturing, at the state by subdivision level. This process, referred to as post-stratification, involved allocation of survey data collected under the ANZSIC 1993 industry classification survey design to the ANZSIC 2006 industry classification basis. Information for both ANZSIC codes was available on the ABS Business Register for units in the Survey in the ten-quarter period from September 2006 to December 2008.
Proration of data was used to extend each historic base series back to the commencement of each ANZSIC 1993 series. The ten quarters of data available measuring the distribution of ANZSIC 1993 series data to the ANZSIC 2006 series was the basis for calculating proportions to be applied in this process.
The next step was to measure the statistical impacts on each Survey of New Capital Expenditure base series, following the changes introduced in the parallel survey for the March and June 2009 quarters. These impacts were the difference between the actual series based on the parallel survey estimates and the survey estimates for the March and June 2009 quarters constructed according to the same parameters as the historical base series. The difference in the level of these two series, termed the statistical impact, has been backcast in the ANZSIC 2006 historical series.
The aim of the above process was to create a time series of estimates on an ANZSIC 2006 basis as continuous and comparable as possible and to maintain, as far as possible, the integrity of the period to period seasonally adjusted movements. This has been achieved at the Australia and broader industry and state levels but the process can result in some volatility at the more detailed state by industry division level.
Estimates to be published in future data releases for March and June 2009 quarters are based on actual data collected in those quarters, under the ANZSIC 2006 design for the Survey of New Capital Expenditure. March and June 2009 quarter estimates will not incorporate backcast data. Backcast estimates in future data releases will be for December quarter 2008 and earlier periods.
The respective impact of individual design changes can not be identified quantitatively for users. Estimates on the basis of ANZSIC 1993 industry classification and net of the various changes described in this paper will not be available for the reference period September quarter 2009 or later.
5 IMPLEMENTATION TIMING
Changes described in this publication to the frame and sample design and outputs for the Survey of New Capital Expenditure will be implemented commencing with the release of September quarter 2009 published outputs for the Survey.
The publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) and associated data products available through www.abs.gov.au are scheduled for release on 26 November 2009.
6 KEY CHANGES TO OUTPUTS
Changes to published outputs for the Survey of New Capital Expenditure are included in the PDF version of this information paper. The affected table (Table 2) has been reproduced, incorporating changes to the format and design, commencing with the September 2009 publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0).
Changes to tables which are published electronically on the ABS web site are detailed for users in the templates which have been provided on the web page for Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) at www.abs.gov.au.
The key changes to capital expenditure outputs are discussed below.
Expansion of published detailed industry outputs
As noted in Section 2, a key change for the Survey of New Capital Expenditure associated with the implementation of ANZSIC 2006 is the more detailed categorisation of data by industry in Table 2 of Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0). The new table structure can be reviewed in the PDF version of this information paper.
The number of industries included in the publication of detailed industries for capital expenditure data has been expanded from nine to twelve to improve the currency and relevance of the data and to better align with other ABS economic collections, notably Business Indicators, Australia (cat. no. 5676.0), published quarterly.
The sample Table 2 shows that the following industries have been extracted from the former published industry category Other Services and will be categorised separately:
The previous published industry category Property and Business Services has been divided into two published industry categories (with Division N, Administrative and Support Services also moving to Other Services):
Other Services in the capital expenditure publication Table 2 will now comprise:
Changes to the publication of Capital Expenditure Realisation Ratio tables
Realisation ratios are the ratio of actual expenditure for a given financial year to each progressive preceding estimate (combining expected expenditure and actual) for the financial year.
Commencing with the September quarter 2009 publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0), the following changes will be implemented for the capital expenditure statistical tables relating to realisation ratios:
A feature article will be included in the September quarter 2009 publication Private New Capital Expenditure and Expected Expenditure (cat. no. 5625.0) which will discuss the determination that a 5-year average realisation ratio is not necessarily a sound basis to project future capital expenditure.
In summary, an average of the previous 5 years’ pattern of capital expenditure realisation of expected expenditure is not considered a sound approach to projecting future capital expenditure when the economy has experienced a significant shift in the economic cycle, with specific impact in an event such as a turning point.
It is considered a better approach to provide the user community with a full history of available capital expenditure realisation ratios and for users to determine which realisation ratios or multiple term average realisation ratio may be suitable for projection of future actual capital expenditure, in a given economic context.
This step has been taken with the inclusion from July 2009 of two additional tables on the ABS web site which feature time series of these statistics back to the first available data. These tables are titled:
These documents will be presented in a new window.