For years we have been bemoaning the lack of quality data provided by the ABS at a regional (SA4) level. In particular the labour market data has been a major gripe of ours. How can we expect State and Local Governments to be making sensible policy decisions when they are fed sub-standard data on which to base those decisions?
At a State and National level the ABS provide Labour Force data on a seasonally adjusted (often taken as the “headline” number) and Trend basis. The ABS are very clear in their message that the Trend series, although not as headline grabbing as the seasonally adjusted, is the preferred measure. We agree.
However, when it comes to the regional data the Labour Force data is presented in its raw, unadjusted, original state. The ABS caution against using the monthly original data (as they should) and instead point users to the 12 month averages (which they also now helpfully provide for those unable to add up and divide by 12). These annual averages are taken as the “official” data for the regions and they are used by the Queensland Government Statistician’s Office, among others, in their regional labour market briefings. The problem here is that these annual averages are all but useless as an indicator about what is actually happening now, or at least in the recent past. Certainly no one would suggest that the annual average data was “good enough” at a State or National level; and certainly no one would accept it if that was presented as the basis for decision making.
The latest ABS Labour Force data for June gives us a good idea of the scale of the problem.
Consider the question “what is happening to employment in Queensland at the moment?”
To answer that question any sensible soul would head straight to the ABS Trend data for June and there they would discover that Trend employment in Queensland sits at 2,391,400 and has risen by 43,300 over the past year.
But if we wished to dig a little deeper and see where those gains were then we’re in trouble.
A look at the ABS annual average data for June tells us that Greater Brisbane has 1,182,900 people employed; a number which is up by 3,500 in the year. The Rest of Queensland employs 1,177,400 and has seen employment fall by 3,600 in the 12 months. How can this be right? Employment at only 2,360,300 (31,100 lower than the Trend) and having fallen by 100 over the year!?
This is why we created the Conus Trend and, despite some reasonable concerns about volatility in the underlying data, we are convinced that it is a major step towards a better understanding of regional labour forces. The Conus Trend for June shows employment in Greater Brisbane at 1,192,700 having risen 11,700 over the year. The Rest of Queensland stands at 1,198,800 and is up 31,700 for the year. That is a very different story to the “official” one told by the ABS annual average but is consistent with the ABS Trend data for Queensland.
At a regional level the differences can also be stark.
- In Cairns the annual average has employment up 6,300 for the year and the unemployment rate at 6.3%. The Conus Trend is much stronger with employment growth of 12,700 and an unemployment rate of 5.3%.
- A similar story emerges in Townsville where the “official” unemployment rate is 9.7% while we estimate it at 7.5%.
- And it’s not all one way; in Fitzroy the annual average shows a loss of 1,500 jobs for the year while the Trend estimates it as a 8,900 fall.
The point, I hope, is clear. The “official” annual average data is in no way reflective of the reality on the ground and therefore cannot be a sensible basis for decision making.
I understand that the ABS has limited resources and feels unable to complete the work required at a SA4 level to provide a series they are happy with. I can appreciate that problem (although at Conus we too have only limited resources and yet have managed to do the work for QLD, NSW and Victoria!…see the full data sets here), but what about taking a preliminary step and providing ABS Trend data at the “Capital and Rest of State” level? Such a move would indicate a genuine commitment from the ABS to the provision of better regional data (and therefore decision making) and provide State and Local governments with a far better idea about what is going on in their Labour Markets.
We’re encouraged that many more people are now using the Conus Trend data that we make available but frankly would prefer to see the ABS step into the breech, even if only in a limited way, and help out the regions.