Covid has a way of making fools of everything, even data. The July unemployment rate released on Thursday of just 4.6% really makes no sense. It certainly is not something worth boasting about, as Josh Frydenberg did when he noted it was the “lowest since 2008” and that it “beat market expectations of 5%”. The situation is so unusual that what usually constitutes a decent measure of the economy has been turned on its head.
In late June-early July, Sydney began lockdowns of varying levels. Not surprisingly, given the bureau of statistics does its monthly labour force survey in the first two weeks of each month, there was a big drop in the amount of work being done in the state.
Hours worked in New South Wales in July fell 7% – some 40.5 million fewer hours in total, the third biggest drop in the history of the state. There was also a 0.9% drop in the number of people employed in NSW and a massive 28% increase in the number of people underemployed.
So it’s not surprising that the state’s unemployment rate was going to be bad – and then it went from a low of 5.1% in June to an even lower rate of 4.5%.
Wait. What?!
That, my friends, is wacky.
The fall in NSW was echoed in the national figures – a drop from 4.9% to 4.6%, despite the biggest state in the nation going into lockdown.
If the graph does not display then click here
When the unemployment rate goes in weird directions usually it is because of a drop in unemployment – not because people have got jobs but because they stopped looking for work.
That certainly occurred in NSW.
The participation rate fell one percentage point to 64.9%. If we added the people who left the labour force in July to the unemployed, the unemployment rate in NSW would instead be 6.0%.
But that is not really the reason NSW’s or Australia’s unemployment figures have taken a turn for the weird.
That participation rate would have been a record high in 2018, so it is not a massive fall, and across Australia the participation rate barely dropped. Even if we took it into account, the national unemployment rate would still be a strong 4.9%.
Neither is it the fault of “one hour of employment counts as being employed”. As I have noted numerous times it’s not a problem, and actually fewer people now work one to nine hours a week than did before the pandemic.
No, the big issue is not working only one hour, but working zero hours.
Each month a bit over 2% of employed people don’t work. Usually about two-thirds of them are on leave. The rest are mostly ill or on maternity and carers leave. They are not a worry.
The big concern is those working zero hours for economic reasons, such as no shifts available, and the amorphous “other reasons”.
Usually, these two categories account for about 7% of all employed working zero hours. In NSW last month they accounted for 44%.
Economist Saul Eslake estimated that across Australia, if you include those working zero hours for economic and other reasons in July, as well as the drop in the participation rate, the “effective” national unemployment rate would have been 8.0%.
I don’t like playing around with the participation rate too much – it goes up and down and is now not that bad. And given around 0.5% of the labour force are always working zero hours due to economic and “other” reasons, I like to discount that amount to highlight the impact of the lockdowns.
Once that is all considered, I suggest Australia’s “effective” unemployment rate rose in July to 6.8% rather than fell to 4.5% and in NSW, rather than a lovely 4.6%, the effective rate was a very sad 9.8%.
The good news is that the experience of the past year has been that once lockdowns ease the labour force recovers quite quickly.
But until we are at a point where lockdowns are no longer an issue we will have to treat economic data with rather a lot of care – and perhaps not get too excited or upset when they don’t seem to make sense.
