An article in the Business Spectator from Adam Carr is suggesting that the labour force data released yesterday is actually much better than it might at first appear. Carr’s argument is that the seasonal adjustment process employed by the ABS (which has certainly been under a cloud recently) is nothing more than “the stuff of fantasy” and “sleight of hand” and that the original (unadjusted) data from the ABS in a few previous months is showing employment growth that “has been one of the strongest, if not the strongest, result on record.”
Carr concludes that “the economy is clearly having no problem absorbing new Australians and growth itself is more than sufficient to see the unemployment rate drop.”
Carr has a fair point when it comes to the issues with the seasonal adjustment and his comments about the strength of the original employment data are, at least in part, accurate. We usually agree with much that Carr writes so were interested to see if his contention about, what we believe are still fairly weak (see our commentary here), employment numbers held up. Unfortunately we don’t believe they do.
For a start we have to accept that the monthly original data is so volatile and subject to seasonal impacts that to consider individual months in isolation is pointless. If we also accept that the ABS seasonal adjustment process is problematic at the moment then a simple way to try and get around the volatility is to consider the original data on a 12 month moving average basis; simply the previous 12 months of data summed and divided by 12. That is what we’ve done in the chart below. We must also realise that the nominal scale of any employment increase needs to be considered in the light of the (growing) size of the Australian population.
We have then considered the difference in monthly population changes with the growth in (12 mma) employment. The resultant graph below shows red columns above the axis when the population is growing faster than employment. As we can see, in times of economic slowdown or recession we get population growth well in excess of employment growth (this is likely to translate into higher unemployment rates, although the scale will depend upon such things as Participation Rates and demographic make-up). You would expect population growth to generally be greater than employment growth since not all people will want to (or be old enough to) take up work.
What we also see is that, despite healthy nominal employment growth (the point that Carr was making), since the GFC we have seen an extended period where population growth has outpaced employment growth by a large extent almost consistently for 8 years. To find a period where employment growth relative to population growth has been so weak we need to go back all the way to the early 1980s. The difference with the period from the mid-90s to the GFC with that since the crisis is marked.
This analysis would suggest that Carr’s conclusion is overly optimistic. His concerns about the ABS seasonal adjusted may be well founded and a consideration of the unadjusted original data warranted but doing so demonstrates weakness in employment growth which is certainly not “more than sufficient to see the unemployment rate drop.”