So, the Australian Bureau of Statistics told us this week, the rate of unemployment fell a tick to 5.6 per cent in July. Trouble is, most people know the official unemployment rate understates the extent of the problem.
What many people don't know, however, is that when you take the rate of unemployment and add the rate of under-employment, which in May took us up to 14.5 per cent, you overstate the extent of the problem.
It's well known by now that the official definition of unemployment is a very narrow one because you only have to do one hour's work in a week to be classed as employed.
A lot of people also know - or think they know - that this amazing definition was introduced by the government some years ago to stop the figures looking so bad.
Labor voters know it was a Coalition government that fudged the figures; Liberal voters know the villain was a Labor government.
Sorry, this is an urban myth. It is just not true. The bureau would never allow any bunch of politicians to fiddle with the definitions it uses.
As it has explained many times, the bureau uses internationally agreed standards to define unemployment, which are set by the International Labour Organisation, part of the United Nations.
And the rule that you're not unemployed if you've worked an hour or more in a week hasn't changed since 1966.
They had to draw the dividing line between unemployed and employed somewhere, and they chose one hour - a choice that was easier to make in the days when almost all the jobs were full-time.
Even today, there'd be very few people actually working just an hour or two a week. Most would work at least one shift of seven or eight hours.
Even so, there's no denying that such a narrow definition understates the extent of joblessness. This is why the bureau also publishes a measure of underemployment.
The underemployed consist of all those people who are working part-time - defined as less than 35 hours a week - but would prefer to be working more hours.
When you take the rate of underemployment and add it to the rate of unemployment (with both unemployment and underemployment expressed as proportions of the labour force) you get what the bureau calls the "labour underutilisation rate", which we can think of as a broader measure of unemployment.
If you look over the years, the rate of unemployment tends to go higher and lower in line with the ups and downs in the business cycle.
You can also see the business cycle reflected in the rate of underemployment, but it has a much clearer underlying upward trend. It was 2.6 per cent in 1978, but 8.3 per cent in November 2015 and 8.8 per cent this May.
Until early 2003, the unemployment rate was higher than the underemployment rate, but since then the underemployment rate has been higher, with a growing gap.
Between February 2015 and this May, the unemployment rate fell by 0.5 percentage points, whereas the underemployment rate rose by 0.3 points.
The underemployment rate is a lot higher for females, 11 per cent, than for males, 6.9 per cent.
It's also greatest among people in lower-skilled occupations and lowest among people in higher-skilled occupations. (Uni students please note.)
Now get this: although workers of all ages suffer underemployment, it's much more a problem for the young. More than a third of the underemployed are aged 15 to 24, and their rate is 18.5 per cent.
But why has the trend rate of underemployment been rising steadily since the late 1970s?
Since underemployment is an affliction of part-time workers, the steady rise in part-time employment over that time - so that it now accounts for about a third of all jobs - does much to explain why there's more part-timers who happen to be saying they'd prefer to be working more hours.
Professor Jeff Borland, of the University of Melbourne, adds that "younger workers appear to have experienced the largest increase in underemployment because they have had the largest growth in part-time employment".
He reminds us that more young people have part-time work because more of them are in full-time education and needing a part-time job.
Adding apples to oranges
But here's my punchline: although the official unemployment rate understates the size of the problem, just adding the underemployment rate goes to the other extreme of exaggerating it.
Why? Because it adds apples to oranges. We worry most about underemployment because we assume it involves people who need full-time jobs but have had to settle for part-time.
It does. But it also includes people who are happy to stay part-time but, even so, would prefer to work an extra shift or maybe just a few more hours.
It doesn't make sense to add people with such a small problem to people with the much bigger problem of needing a full-time job but not being able to find one, as though they were similar.
Remember, too, that almost a third of the people included in the official unemployment rate are looking only for part-time work.
This is why, if you search very deep on the bureau's website (clue: catalogue no. 6291.0.55.003, table 23b) you find that, as well as just counting heads, it also does a more accurate measure of underemployment that counts the hours people are looking for - meaning part-timers needing a full-time job count for a lot more than those just wanting a few more hours.
This "volume" measure shows that, in May, the underemployment rate was 3.2 per cent of all the potential hours the whole labour force could work, and the unemployment rate was 4.3 per cent, giving an hours-based measure of labour underutilisation of 7.5 per cent.
Which is closer to the truth of the matter.
Ross Gittins is the Herald's economics editor.