When an economy gets hit we need to work out first when we have hit the bottom, and then determine when we can say the recovery is over. In the current crisis, the bottom is determined by the lockdown restrictions – most of which are at least less than they were in April (even in Victoria). But when we can say we are back to normal will depend somewhat on what we decide that normal will be.

The latest unemployment figures, out last week, highlighted how crazy everything is right now. The government initially suggested the unemployment rate would peak at 10%, but now Treasury thinks 8% might be more realistic.

That would seem to be good news, except the unemployment rate has been rendered a rather ineffectual measure because so many people have left the labour force that the current “real” rate of unemployment is closer to 11.3%.

So we could argue that we will know the crisis has peaked when the unemployment rate starts going down, but ironically it will probably rise when things get better because more people will come back into the labour force and be counted as “unemployed”.

We could look at the participation rate, which counts both employed and unemployed. Certainly that needs to rise before we can say we have hit the bottom. But it is less helpful for determining when we have recovered.

Just prior to the virus, the participation rate was at a record level, but that was mostly due to a large rise in women entering the workforce.

As it is, the massive drop since March has seen women’s participation fall back to 2007 levels and overall participation go back to 2000 levels, while men’s participation is now at lows never before encountered:

We could choose to look just at the growth of employment and hours worked:

Over the past decade or so, both measures have averaged growth around 1.6%, but over the past three years we have seen very strong employment – a touch over 2% growth. So perhaps when we return to that level of growth things will be back to normal.

But we need to remember that current growth figures hide past damage. Even if employment began growing right now at an annual rate of 2% it would take nearly three years to get back to where we were in March:

But of course our population would be bigger then, so we would still be worse off. To take that into account we need to look at the employment rate – the percentage of all adults who are employed.

This has taken a massive hit – falling from 62.6% to 58.4% in two months.

But even here we need to be careful when talking about returning to normal, for this is affected by the same factors that drove the generally falling male participation rate.

The history of past recessions shows that while it can take many years for the employment rate to recover, the level of people working fulltime never does – and most of that work is done by men:

Prior to the 1990s recession, 59.7% of all adults were employed, and 47.1% were working full-time. By the time the employment rate recovered, a decade later in 2000, just 43.8% of all adults were working fulltime.

Similarly, while at the start of this year total employment had nearly returned to pre-GFC levels of total employment, the percentage of adults working fulltime had dropped by around 2.5%pts.

This is why it is good to look at hours worked.

Prior to the virus the percentage of adults working was just a touch below the pre-GFC level but the level of hours worked per capita had fallen more than 4%.

Now it is down 15%:

It may be that the pre-GFC level of 89.9 hours being worked per month was unsustainable. Clearly everything had to be going right – a massive mining boom and a strong global economy driven by China.

Over the past 40 years the average level worked per month per capita (which includes everyone over 15, whether in the labour force or not) is around 85.7 hours:

Coincidentally, that was where we were just before the virus hit.

And so while the unemployment rate may bounce around, and the participation rate zoom back, and we see retail sales going up and down faster than a rollercoaster, focusing on the hours worked per capita will be a strong guide for letting us know when we’re back to normal. Until we see at least per capita monthly hours worked back to 85.7 hours, talk about slowing down stimulus spending will be greatly premature.

But even that will not fully tell us what the jobs market will be like at that time. How many of us will be working fulltime? What about wages growth, and employment conditions? Will the work have shifted strongly towards older people, with the youth left behind?

Whatever the measure we use, we can hope the bottom has been reached, but the path back to “normal”, and what that normal will be, is yet to be determined.

• Greg Jericho writes on economics for Guardian Australia