Warnings about migration data were strong, but could have been stronger: ABS (News Backgrounder 11/9/2003)

The Australian Bureau of Statistics notes the publication of several critical media articles on 8 September concerning official migration statistics.

These articles were prompted by the release of research by demographers from the Australian Centre for Population Research, Australian National University.

Since becoming aware of a data problem with migration statistics late last year, the ABS took what it believed to be sufficiently strong steps to alert users of these statistics to potential problems about data accuracy.

These steps included three published warnings and a demographic working paper about the problem of correctly counting travellers who change their intentions.

Given the strength of feedback from commentators the ABS concedes that the warnings could have been more prominent and less technical in the interests of a well-informed community.

ABS has been evaluating alternative methods to overcome the data deficiencies, and updated migration statistics are expected to be available by the end of the year.

The ABS welcomes the useful work by the ANU in this complex area of demographic statistics, and will take account of this work in its evaluation. Currently the ABS has some concern that the ANU figures are too high.


The ABS has been researching a new method to estimate for travellers who change their travel intentions (known as category jumping).

For overseas visitors to Australia, and Australian residents leaving Australia, their intended duration of stay or absence may change over time. The changes in travel intentions determine how they are treated in compiling Australia's migration and population estimates.

Given that it has become necessary to change the methodology used in the past, ABS research is considering alternative options for adjustment. That research, which uses different methods to the ANU study, has not yet been finalised. Revised official estimates are not therefore available for release at this time.

ABS provided three warnings to statistics users, including the following on 20 March 2003:

    "Category jumping for the period from September Quarter 1997 to June Quarter 2001 have been set to zero and are final. Data for September Quarter 2001 to September Quarter 2002 have also been set to zero but will be revised when additional information becomes available. This was necessitated by the recent detection of a deficiency in the current measurement of migration category jumping. For further information please see Demography Working Paper 2003/1 - Estimated Resident Population and Measurement of Category Jumping"
Page 2, Australian Demographic Statistics, September Quarter 2002 (ABS cat. no. 3101.0)

ABS will be evaluating the ANU research findings along with the ABS methods under development and expect to apply revisions to the level of net overseas migration by the end of 2003. Additional work is needed before any revisions can be applied to the ABS migration and population statistics.

The ABS recognises the importance of accurate migration statistics for decision making.

Accordingly, ABS has been openly advising users of the short term data quality caveats while new methods are being developed to overcome recently detected source data limitations.

For further clarification please contact: Patrick Corr Tel (02) 6252 6411, Mobile 0419 513 569.