Monday, September 22, 2014

Painting by numbers

I haven't posted anything yet on last weeks regional employment data which has been covered by Conus: The good news for jobs continues in Cairns

Pete has also updated and made available the Conus Trend employment and unemployment rate for the Queensland regions. This is particularly valuable in being able to provide a more sound basis of comparison between regions for the notoriously unreliable monthly data from the ABS.

The ABS provide data for the Queensland SA4 regions and also a break this into 'Greater Brisbane' and 'Rest of Qld'. What I have looked at is the Y-o-Y change in the Conus Trend from August 2013 to August 2014. In this period there were 49,300 jobs total created in Qld. 22,900 in Greater Brisbane and 26,400 in Rest of Qld.

Fair enough, but when we look at the SA4 breakdown then 21,100 of those regional jobs (80%) were on the Gold Coast. If one were to take the liberty to rearrange the statistical geography and include both Gold and Sunshine Coasts with Greater Brisbane in 'SEQ' then there would have been no net jobs gain at all over the year outside SEQ.

If we rank the Queensland regions on employment growth this is how it looks:

Gold Coast is also the largest region by population here. The validity of the 7.7% growth on the Gold Coast looks a bit of a stretch to me even on the trend numbers. (Note: wouldn't place much relevance there on 'Outback' given the relatively small population and sample size with a geography that covers a huge area of inland Qld including Cape York.)

However, within Greater Brisbane there are also divergences as wide as plus 9.7% employment growth in the Inner City to minus 9.2% in Logan-Beaudesert.

Should be careful here, as Ricardian Ambivalence always used to warn the methodology of the labour force survey is to derive an unemployment rate and employment is a secondary estimate based on further assumptions such as population growth. So the significant point to note is that even with such strong employment growth the unemployment rate on the Gold Coast increased:

This is why some of us keep banging on about participation rates and the need to consider the full range of indicators and not just a selected component. To play with some numbers I have combined the two above to correlate employment growth with the unemployment rate:

I wont pretend that trend line has any significance. Despite that there doesn't appear to be any meaningful correlation between employment growth and the unemployment rates across the regions? Cairns is one of only three to show improvement in both (bottom right quadrant). Fitzroy is particularly interesting stand-out there given the association with resource industries.

Stephen Koukoulas tweeted in response to the recently contentious national stats that there was a danger of a policy mistake if reliability of the data was not improved. The ABS has also been subject to recent funding cuts. I guess we just have to trust that the RBA understand the problems and complexities.

At a regional level it is now Townsville which is angsting hard based on the unemployment stats. I appreciate that Townsville may be soft and the outlook weaker than other regions but suspect there is a danger in being hostage to this employment data, as I think previously happened in Cairns: Townsville jobs disappear amid 'perfect storm' of economic problems

I remain content to be circumspect on all the labour force data currently although recognise that such confused circumstances could correlate with a turning point.

No comments:

Post a Comment