How do I report employment stats responsibly?
The worst labour figures dashboard in history.
I’m trying to decide if my obsession with labour force figures comes from a genuine curiosity and appreciation of their importance or the ease with which one finds said numbers.
I’m thinking that it’s the former, given that I wouldn’t be asking the question in the title if I weren’t somewhat genuine in my concern.
Nearly a decade ago I sat in an economics lecture and nodded along as the person at the front complained that the media were daft if they reported volume figures for unemployment and should have been reporting the rate instead.
I didn’t fully appreciate the need for that choice until this week when I started charting the buggers.
Not only did I realise that the volume figures told me nothing about the real size of the problem (10,000 new unemployed means nothing if the workforce population just blew out by a million), but that even the unemployment rate was useless as a headline figure; the compression of data just squeezes all the insight out of it.
So, like the good baby-journalist that I am, I dug deeper.
This wasn’t hard, underemployment and underutilisation figures have been reported by the ABS since the late 70s and come out literally right next to the data that usually gets reported.
ABS: 1 Media: 0
Yesterday I included underutilisation in my labour force data dashboard (published on all good APN news sites near you) but still felt like I wasn’t able to understand the real shape of employment in Australia.
There are near-constant cries about unemployment on our news sites and I couldn’t really reconcile that with the headline figure. 5.4%? Are you kidding? That’s a decent holiday retail season away from us having problems with demand-pull inflation! (citation needed).
My brother Anthony reminded me of just how exceedingly bullshit the ABS’s definition of employment is. 1 hour worked for pay each week or 14 hours at home or on a farm unpaid and you’re employed.
So how many people are in the group between unemployed and able to put food on the table. I argued that this was captured by the underemployment rate but Anthony pressed harder, that there’s no way to tell the difference between someone who works one hour and wants more employment and someone who works 25 hours and can feed themselves, but wants more employment.
Back to the stats I went and found the figures for hours worked. I haven’t really gotten to know them intimately yet but they seem to be what they say they are. I’ll be including them as well.
Unfortunately there’s no way I can see that the readily available figures can tell me who is underemployed AND works how many hours.
If someone works an hour a week and is fine with that because their partner maintains the bacon supply, I don’t want to capture them in the numbers for people who are being missed by the headline figures.
Initial investigations suggest that I’ll need to contact the ABS and get these figures specifically.
Any advice on the matter would be greatly appreciated.
I get the feeling that this isn’t the only valid criticism of the reportage of labour force figures in Australia. If you know of any others, please leave a comment in the comment section below.
Good post!
But I suspect you are over thinking it a little bit!
The way I see it is that you are having an issue with the purpose of the data, not so much the data itself.
The unemployment rate is an abstraction of the impacts of unemployment/underemployment etc. Economists understand the limits of the statistic and know how to apply it to estimate things like consumer confidence, spending and saving rates etc.
What you seem to be puzzling about is how this figure can be misleading to the lay person because it excludes a lot of people that the lay person would consider to be unemployed. Worse still it excludes people who would describe themselves as unemployed.
Which on the surface seems bad but it gets worse. Unemployment is experienced locally, you can have areas of the country/state with massive unemployment and areas with practically no unemployment and people only experience this locally. Geographic is one thing but unemployment also varies by demographic. A 20 year old reader may scoff at a 5% unemployment figure saying 1 in 5 of their friends is unemployed. A 50 year old may say that 5% seems high saying that 1 in 50 of their friends is unemployed. Not just age but also gender, race etc.
Which is actually a really good thing to you. Your readership is simply too wide to provide a meaningful portrayal of the unemployment rate across the wide lived experiences of your readers. Just throw up the numbers, volume and percentage swing and if any of your readers want to question how the figures are calculated they can ask the ABS.
I think we may have reached a whole new level of cynicism with regards to reporting unemployment. At the very least we can be proud of that.
I agree that the stats as they’re delivered by the ABS are, IMHO, very well tuned for the modelling that economists use and, in the end, it’s more important that economists and policy makers understand what’s going on than our readers.
Before I go on, I’d like to explain that for others who may be reading. It’s more important for an economist to have the stats tuned for them than tuned for a reader simply because it’s the economist who’s tasked with fixing the problem. All a reader can really do is, at best, vote sideways or complain. An economist knows that at x level of unemployment and y level of underutilisation, z must be done to achieve .
Ideally, any level of unemployment is bad, but at certain levels approaching zero the trick isn’t to ‘create more jobs’ but to finesse demand in such a way that you don’t cause inflation to spike.
Back to why I’m looking to report a slightly more detailed and customer-facing set of figures.
Employment level reporting can’t be extricated from voting decisions. It’s a foolish person who sees employment figures plummet and doesn’t factor that into their voting decision. The current degree of abstraction in how the media reports employment isn’t appropriate for that decision process.
You can say almost anything about the employment situation on our comments section and there’s really no way to verify how accurate that is. I can say “Unemployment is close to full employment, well done Labor” while someone else can say “People are struggling to make ends meet, fuck you Labor” and both of us would be fairly close to reasonable in our assertions.
I’m seeing labour force figures as being part of the big three, money supply, employment, and inflation. I’m just trying to make sure there’s sophisticated representation of the employment part of that equation.