1352.0.55.034 - Research Paper: Methodological Issues Associated with Using BAS Data for Improving Sample Design and Estimation for Business Surveys (Methodology Advisory Committee), Nov 2000  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 10/11/2000  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All
  • About this Release

About this Release

Under the new taxation system, there will be a quarterly census of businesses in the form of Business Activity Statements (BAS) returned to the Australian Taxation Office. The data collected on these statements will be made available to the Australian Bureau of Statistics, which has made a corporate decision to use the data to effect gains in terms of reduced sample size, reduced provider load and reduced survey costs. The ways in which these gains may be achieved have been identified but only some are relevant to this paper. One is to redesign business surveys by replacing our existing size stratification with a new one based on BAS variables. Another is to improve our estimation procedures by introducing the use of generalised regression estimation with BAS variables as auxiliaries. The third is to use BAS data as a substitute for currently collected survey data.

In this paper I examine each of these ways of achieving gains by considering an approach that will require the minimum amount of change to our current procedures and then a more comprehensive approach. In the case of estimation the minimum approach is to use ratio estimation but with new auxiliary variables while the more comprehensive approach involves implementing generalised regression. Finally, I discuss how the three ways of achieving gains will interact. Some of the issues that have arisen are theoretical and some are practical. They are listed in the Appendix at the end of the paper.