We have been highlighting for some time the fact that the Queensland labour market has benefitted from the positive health outcomes in the State (see here for the most recent October data). This analysis has not been universally agreed with and some commentators have pointed at alternative data sets, such as Payrolls or JobSeekers, to suggest that in fact the State’s labour market remains weak. The Queensland Treasurer, Cameron Dick, has been spruiking the fact that the ABS data suggested October QLD employment was a touch higher than in March and that, on that basis, the State had recovered all its lost COVID jobs. While this may be identified as a rather ‘bold’ claim by some it is one that the data, at least on face value, supports.
A common strand of counter-argument to that of the strong Queensland labour market has run along the lines of “If things are so much better in Queensland then why do we still have one of the highest unemployment rates in the country?”. Australia’s unemployment rate in October stood at 7.0% while Queensland’s was at 7.7%. At first glance that is indeed a fair critique, although I hope to demonstrate below the weakness in this argument.
We have heard a lot about the distorting effects that the JobKeeper payments, and the increased JobSeeker payments, have been having on the labour market and many, including the ABS, have discussed the calculation of an ‘effective unemployment rate’. This rate is adjusted so as to account for all those who are working zero hours and for those who have left the labour force completely since March.
The adjustment is done by including those working zero hours (who are currently counted as employed by the Labour Force Survey) as unemployed. Additionally the labour force is estimated as it would have been had the participation rate as at March 2020 been current now. The difference between that assumed labour force and the actual labour force is calculated. In Australia, where the participation rate fell sharply, this has been a reduction in the labour force. The ‘effective unemployment rate’ assumes those who have left the labour force would instead also be counted as unemployed.
We see that the Australian March unemployment rate of 5.2% (which would already have been 5.8% had we included zero hour workers) was based on a participation rate of 65.9. Since then, as the number of zero hours workers increased sharply and the labour force fell sharply, the ‘effective unemployment rate’ got as high as 15.8% in April before starting to decline. It now sits at 8.1%; or 2.9ppts above the official unemployment rate in March and 1.1 ppts above the official rate today.
When we consider the Queensland data we see that the March unemployment rate of 5.6% (which would have been 6.1% if we had included zero hour workers) with a participation rate of 65.5. Since then the ‘effective rate’ peaked at 16.7% in May but has now sharply declined as the participation rate moves above where it sat in March (and the labour force adjustment therefore subtracts people from the count of those unemployed). Based on an estimate of 19,900 zero hours workers in October (this is based on the same percentage reduction in QLD as we saw across the nation from Sept; actual data will be released this Thursday) we see that Queensland’s ‘effective unemployment rate’ is only 6.8%; well below the national ‘effective rate’, only 1.2ppts above its level in March and 0.9 ppts below the State’s official rate.
What this suggests is that the official unemployment rate in Queensland is as high as it is not because of a particularly weak labour market, but rather because the better health outcomes have resulted in a strong participation response.
On this same basis of calculation, and estimation of zero hours workers, the NSW ‘effective rate’ in Oct will likely be 6.5-7.0% (up 1.6-2.1 ppts from March and similar to the official rate of 6.5%) and Victoria will still be around 12.0% (up about 6.8 ppts from March and still about 4.5 ppts above the official rate of 7.4%).
We can also see that, with the current exception of Victoria, the distorting effects of the JobKeeper payments has been largely unwound with the actual and ‘effective’ unemployment rates converging.