Thursday, March 12, 2009

Lies, Statistics and Employment Data

Earlier today the Australian Bureau of Statistics published employment data relating to the month of February. Recent budget cut-backs however makes the data unreliable.

Last month in an interview conducted on ABC's PM program the February release of employment data was critically analysed. The critical analysis was a direct result of figures indicating the (full-time) employment and participation rate had actually increased. These results startled many economists as it contradicted all the information published leading up to the release. Specifically, it showed the number of employed had increased by 1200.

How can this be possible?

The explanation of this result can be attributed to statistics. More specifically, the lack of (federal Government) appreciation for robust statistical analysis. Closer examination revealed that employment may have decreased by as much as 60 000, or possibly gained nearly 62 000. The reason for this imprecision is sampling error. The reason for the sampling error was a cut to the budget of 24% which meant the ABS slashed the size of the labour survey. The coverage is now nearly half of what it was the same time last year (down to 0.24 from 0.45)

Does it matter?

Yes, for three reasons:
  1. Employment data is one of the most featured variables in gauging economic well being.
  2. Without reliable employment data it is much harder to critically assess Government policy. That is transparency and accountability suffers.
  3. The information is disseminated all over the globe so that governments and businesses can use this information in decision making.
Therefore, it is absolutely imperative that when employment data is published it is reliable. The current arrangements mean it is not. I believe this damages the reputation of the Australian economy and possibly even raises questions about its integrity.

No comments:

Post a Comment