FORTHCOMING CHANGES


RECENT DEVELOPMENTS

A note in the April 2008 issue of this publication advised of significant changes to the Retail Business Survey occurring from the July 2008 reference month. The changes include an approximate two-thirds reduction in the sample size used to create monthly estimates. This note summarises the key changes and provides an update on three key aspects of the redesigned survey - sample design and standard errors, monthly output and quarterly output.


KEY CHANGES TO THE RETAIL BUSINESS SURVEY

The key changes to be made to the Retail Business Survey and this publication are:

• the aim of the survey will be to produce a quarterly series with similar detail (apart from the industries mentioned below which are now excluded) to that released in the current monthly publication whilst maintaining a broad level monthly series
• a new sampling methodology will be adopted whereby each sampled business is only required to provide data for one month in each quarter
• commentary in the monthly publication will focus on the trend series
• there will be a new separate quarterly publication which will include commentary on seasonally adjusted series
• Hotels and licensed clubs and Selected service industries will be excluded from the survey.

There will be increased volatility in month to month movements in the original and seasonally adjusted series because of the smaller sample size and the rotation effect of having a different third of the sample reporting each month.


REVISED SAMPLE DESIGN AND STANDARD ERRORS

The sample allocation for the redesigned survey has been completed. The redesigned survey will include 3200 businesses each quarter composed of 500 large businesses which are surveyed each month and about 2700 sampled businesses which are surveyed once each quarter (900 sample sector businesses included each month). By comparison, after excluding Hotels and licensed clubs and Selected service industries, which will be excluded from the redesigned survey, the sample size for the current survey design is 3415 businesses each month composed of 640 large businesses and 2775 sampled businesses.

Estimates have been modelled from the current sample, using the reduced scope, sample allocation and estimation methodology that will be used for the redesigned survey. The tables below compare the average relative standard errors resulting from the current survey design (including the full scope for States and Australia) with the relative standard errors based on the estimates modelled on the new survey design. The latest modelled relative standard errors are lower than those included in the note in the April issue.

1. Retail relative standard errors, State
	Average RSEs from current design(a)	Modelled RSEs from new design
	Monthly	Monthly	Quarterly
State	%	%	%
New South Wales	1.5	2.1	1.4
Victoria	1.5	2.2	1.5
Queensland	1.6	2.5	1.7
South Australia	1.7	2.2	1.5
Western Australia	1.7	2.6	1.7
Tasmania	2.6	3.0	1.9
Northern Territory	2.0	2.3	1.6
Australian Capital Territory	1.7	2.7	1.9
Australia	0.7	1.1	0.8
(a) Average of RSEs achieved by the current design from the July 2007 to March 2008 months.

2. Retail relative standard errors, Publication industry
	Average RSEs from current design(a)	Modelled RSEs from new design
	Monthly	Monthly	Quarterly
Publication industry	%	%	%
Food retailing	0.9	1.1	0.8
Department stores	0.0	0.0	0.0
Clothing and soft good retailing	2.6	3.6	2.5
Household good retailing	1.8	2.6	1.8
Recreational good retailing	4.6	6.0	4.0
Other retailing	3.6	4.7	3.3
Australia	0.7	1.1	0.8
(a) Average of RSEs achieved by the current design from the July 2007 to March 2008 months.

MONTHLY OUTPUT

Monthly output from the redesigned Retail Business Survey will be considerably less detailed than that currently available monthly. State by industry data will not be available monthly and industry detail will only be available at the broader group level.

Only trend data (level, percentage change from previous month and percentage change from same month in the previous year) will be presented in the tables included in the publication. The original and seasonally adjusted estimates will have increased volatility from the new sampling methodology and are considered of limited use for measuring month to month movements. However, for analytical purposes seasonally adjusted and original data will be made available in the time series data spreadsheets provided with this publication on the ABS website.

Trend data will be provided for Australia, industry groups and State totals. Industry groups to be included in the publication are Food retailing (excluding takeaway food retailing), Cafes restaurants and takeaway food services, Department stores, Clothing and soft good retailing, Household good retailing and Other retailing (including the Recreational good retailing industry which is currently shown separately in this publication).

The following series will be available from the Details tab of the publication to assist users in their understanding of the trend series and enable them to undertake their own detailed analysis:

• Retail turnover level by industry group - original, seasonally adjusted and trend
• Retail turnover percentage change from previous month by industry group - original, seasonally adjusted and trend
• Retail turnover percentage change from same month in the previous year by industry group - original, seasonally adjusted and trend
• Retail turnover relative standard errors by industry group - original
• Retail turnover level by state - original, seasonally adjusted and trend
• Retail turnover percentage change from previous month by state - original, seasonally adjusted and trend
• Retail turnover percentage change from same month in the previous year by state - original, seasonally adjusted and trend
• Retail turnover relative standard error by state - original
• Retail turnover completely enumerated sector total level - original, seasonally adjusted and trend.

Appropriate caveats will be included with the above series advising that care needs to be exercised when using the original and seasonally adjusted series.


QUARTERLY OUTPUT

Analysis is continuing on the level of detail that will be released quarterly. Subject to this analysis it is expected that the level of detail provided will be similar to that currently released in this publication (apart from the industries excluded from the sample redesign). Commentary will continue to emphasise trend data but will include commentary on seasonally adjusted series. The broad industry data will be the same as that to be released on a monthly basis.