REVISIONS TO BUILDING APPROVAL STATISTICS

The Australian Bureau of Statistics (ABS) collects building approval records from approving authorities across Australia. The outputs from the Building Approvals collection are released monthly in the Building Approvals, Australia (cat no. 8731.0) publication. In addition to informing monetary policy and high level macroeconomic research, these statistics underpin the production of fine geographical outputs such as regional economic indicators and estimated resident population. To ensure building approval statistics are as accurate as possible, approval records provided to the ABS for any reference month may be revised in later publications due to corrections made by approving authorities or the late provision of approval records. The ABS also conducts large scale reviews of approving authorities to identify systemic issues resulting in the misclassification, late provision or exclusion of approval records in their regular submissions. This article will outline the sources of revisions in greater detail and explain the principles that guide the timing and magnitude of revisions to building approval statistics.


SOURCES OF REVISIONS

The major drivers of revisions to building approval statistics include:

• Corrections to approval records
  • Approving authorities may advise of corrections or other changes to previously supplied data.
  • Approval records may also be altered by the ABS directly, based on confrontation with other sources of data.

• Identification of additional approval records
  • Additional approval records can be identified on advice from the approving authority itself or through large scale reviews of data from approving authorities conducted by the ABS (see next section).
  • Additional approval records may also be identified through responses to the Building Activity Survey and interrogation of other (independent) data sources by the ABS.

LARGE SCALE REVIEWS

The ABS has previously conducted large scale reviews of building approvals data by requesting that all approving authorities resubmit all approval records for a given period. The last such review was conducted in 2012 and examined all approvals data back to July 2001 (refer to the January 2013 issue of the Building Approvals, Australia (cat no. 8731.0) publication). Recently, the ABS commenced a modified review program, examining data submitted from July 2012-June 2015. Data submitted as part of this review will be compared against the original monthly submissions from approving authorities. From this analysis, the ABS will revise previous building approval statistics and provide feedback to approving authorities on the quality of their submissions.

The ABS intends to conduct similar reviews each year from 2016 onwards, focussing on the previous year's submissions from each approving authority.


PRINCIPLES OF REVISIONS

The following principles govern revisions to building approval statistics. Revisions are made:

• where an approval record has a significant impact on the aggregate estimates (including fine level geographic outputs) to which it contributes; or
• where an approval record is one of multiple records that, as a collective, have a significant impact on the aggregate estimates (including fine level geographic outputs) to which they contribute.

PUBLICATION OF REVISIONS

Revisions made to reference months within the current or previous financial years are published as soon as they are identified, in the next release of the Building Approvals publication. Significant revisions made to reference months prior to the previous financial year are published once per year in the next January release of the Building Approvals publication.

As part of the review process, requests for resubmission of data are spread evenly across the year in order to avoid spikes in revisions in a particular reference month. A higher number of revisions may be observed in the January 2016 and January 2017 issues, as the current review period includes data back to July 2012 (beyond the previous financial year). However, the majority of revisions in the recent and upcoming reviews are expected to occur within the current and previous financial years, and will therefore be published directly after they are identified, in the next release of the Building Approvals publication.

The impact of revisions to the total number of dwelling units approved at state level are presented in the ‘REVISIONS THIS MONTH’ section on page 2 of each publication. From this publication onwards, these numbers will also reflect revisions resulting from the recent and upcoming reviews.

Subsequent impacts of revisions can be seen in the Building Activity Collection and in the National Accounts due to the relationship of these collections with the Building Approvals Collection.