Recent Australian unemployment rates didn't seem to make sense. Here's what is going on

The August labour force figures surprised everyone – unemployment was expected to rise given the Melbourne lockdown, and yet it fell. And while this is good news, a dig into the figures reveals that it needs to be tempered with the realisation that we remain deeply within a recession.

The unemployment rate falling in August from 7.5% to 6.8% (the second biggest monthly fall ever recorded) brought to mind the old line that if you understand what is going on, you clearly have not been paying attention.

It really did not make sense.

Related: Australia’s unemployment figures are surprisingly good but the recovery is slowing | Greg Jericho

In August, the Melbourne CBD moved into stage 4 lockdowns and most economists believed this would surely see unemployment rise. And yet not only did the national unemployment rate fall, but the rise in Victoria was small – up to 7.1% from 6.8% in July. By comparison in April and May it rose 0.8% pts and then 0.9%pts.

To work out what is going on we need to go into the weeds a bit. I know this might feel a bit nerdy economics, but stay with me – it’s not that bad.

The first thing is we need to realise that the figures we are talking about are “seasonally adjusted”. Each month the ABS (like every other statistical agency in the world) adjusts figures to take into account the month/season we are in.

In reality the ABS estimates employment rose by just 44,500 in August, and yet you would have read in all news reports that it grew by 110,000. That’s because that last figure is the “seasonally adjusted” one:

That figure is not made up – it is a statistical calculation that takes into account that for example in January employment always drops, and it really isn’t useful to think that suddenly there is no work in that month each year.

The problem is that seasonally adjusted works when things go as usual, and clearly they are not doing that right now. That means the adjustment can get a bit screwy and subject to revision.

Also, right now there is a fair bit of trouble getting a representative sample given jobseeker/jobkeeper have been tweaked and are subject to different requirements depending on what industry you work in (for example higher education).

That doesn’t mean we should not trust the figures, merely that we probably should not live or die by each month’s result.

There was also a lot of coverage about the increase in sole traders suddenly getting work – up 50,200 in August. The ABS even put out a release noting the “Strong employment growth for non-employees”

The thing is, August is always a strong month for non-employees and a bad one for employees:

This month was a bit stronger than usual for non-employees, but very much better than usual for employees:

So it seems that the job growth was not just about sole traders suddenly saying they were working to ensure they met requirement for government benefits.

Because while employment growth grew 0.9% in August and hours worked increased just 0.1% a big reason for that was Victoria. If we exclude Victoria, employment and hours worked increased by about the same amount:

So all is well? All we have to do is not count Victoria?

Well, aside from the fact that excluding Victoria is a bit like Donald Trump saying the US is doing well on Covid as long as you don’t count “the blue states”, the reality is we are in a deep hole even without Victoria.

The monthly hours worked per capita both with and without Victoria being counted remain lower than at the worst point of the 1990s recession:

Think on that – we have “recovered” to now being just slightly worse than it ever was during the longest recession of the post-WWII era.

And if we look at the full-time employment rate of prime-aged working men (which always gets smashed during a recession) we see that there has been almost no improvement from the nadir of June:

And to cap it all off, we need to realise that 18% of the labour force are still either unemployed or underemployed. If we take out Victoria that just drops to 17% – still as bad as has ever been recorded.

When we use this underutilisation rate to measure the level of recessions in a manner similar to how American economist Claudia Sahm uses unemployment, it is clear we are still as deep within a recession as we have ever been:

So yes, the August unemployment figures were better than expected, but given how crazy things are at the moment, it is best to wait a month or two to see if that was a one-off or a sign of things to come.

But even if the good news holds, do not start thinking we’ve nearly recovered. We have not.