What the Unemployment Number Actually Measures

Quarterly Labour Force Survey day, and two different unemployment rates go round for the same three months. One low, one high, sometimes a lot higher. The usual reaction is that somebody must be spinning something.

Nobody is. Both figures are real. They are counting different people, and the difference is worth understanding before you argue with either of them.

The narrow definition

The official rate, the narrow one, counts people who are out of work, available for work, and who actively looked for work during a set recent window.

That last clause does all the damage. If you did not search in the window, you are not unemployed on this measure. You are outside the labour force entirely, which is a strange thing to call somebody who wants a job.

It is not a local quirk, for what it is worth. The definition is the standard international one, which is exactly why we use it: it lets you put South Africa next to Brazil or Spain and compare like with like.

The expanded definition

The expanded rate puts the searchers back in. Specifically it adds discouraged work-seekers, meaning people who want work and could take it but have stopped looking, usually because they have concluded there is nothing to find, or because looking costs money they do not have.

Where unemployment is very high, and where the nearest realistic job is an expensive taxi ride away, that group gets large. Which is why the expanded rate can sit well above the narrow one.

And here is the part most coverage skips: the size of the gap is itself data. A widening gap is a statement about how many people have stopped trying.

The gap between the two rates is not noise. It is a measure of discouragement.

So which one is right?

Both, because they are answers to different questions.

Want to know how many people are actively competing for jobs right now? The narrow rate is closer to that. Want to know how many people want work and are not getting it, including everyone who has stopped asking? The expanded rate covers more of it.

Choosing one is fine. Quoting one without saying which is the problem, and it happens constantly. A minister announcing progress reaches for the lower figure. A critic reaches for the higher. Neither has lied about a number. Both have quietly declined to mention which number they picked.

How to read the next release

Before you react to any unemployment figure, ask which rate it is. Then go and find the other one.

Watch the gap in particular. A falling narrow rate alongside a widening gap is not obviously good news: it can mean people moved out of active searching and into giving up, which flatters the headline while the underlying position gets worse.

StatsSA publishes both, every quarter, in the same release. It takes about a minute to check. What would change in the next election cycle if the second number got quoted as often as the first?