|Page tools: Print Page Print All RSS Search this Product|
New QBIS sample design
Since its inception in 2001, the Quarterly Business Indicators Survey (QBIS) has been the main vehicle for providing economy-wide data to National Accounts Branch (NAB) to support the preparation of the quarterly national accounts. The survey collects values of national sales, wages, inventories and profits and of state sales and wages as well. All industries other than Agriculture are surveyed, though the data collected vary slightly with industry and size of business.
QBIS is currently undergoing a substantial re-design for the twofold purpose of implementing the new ANZSIC06 industry classification and to achieve the goals of the first stage of the QEWS Phase 3 Project given its current formulation as approved in October 2007.
The first stage of this project was originally planned to consist of:
Implementation of such a significant change presents a number of challenges, not the least of which is to determine the impact on both levels and movements. The ANZSIC change and the inclusion of non-employers both affect the level of the series, but the seasonality is also likely to change, since the seasonal behaviour of the non-employer contribution is expected to be different from that of employers. To capture both impacts, a two-stage implementation strategy is being planned.
A parallel run will be conducted in March and June quarters of 2009 in which two samples will be produced - one from the current design and one from the new design. The sample selected on the current design will be used to produce current outputs - ANZSIC93 employer-only estimates. The sample from the new design will also produce employer-only estimates but on an ANZSIC06 basis. The difference between the two provides a measure of the impact of the ANZSIC change. ANZSIC06 estimates will be published for the first time for the September 2009 quarter.
Expansion of scope without a corresponding increase in sample size always presents a challenge for sample design and the present case was no exception, especially in view of the requirement that the design be capable of delivering acceptable estimates for the employing sub-population as well as for the total (employing plus non-employing) population. In principle, these requirements could be met by stratifying employers and non-employers separately. However, this would produce too many small strata and thus too large a sample. A compromise has therefore been reached in which there are separate strata for non-employers in those industry-by-size groups in which they are a significant proportion of the population. In all other cases (the majority), employers and non-employers are stratified together.
For further information, please contact Jos Beunen on (02) 9268 4647 or firstname.lastname@example.org.
These documents will be presented in a new window.