I have always found taxi drivers to be enthusiastic economic commentators.
This week, I had a great chat with one driver who told me his plans to take a job as a security guard at a mine in Western Australia.
He didn't fancy driving trucks but, as it turns out, there are plenty of well-paid jobs out west, from cleaners to guards, if you can bear the boredom.
A survey released this week by the Bureau of Statistics confirms the finding. Of the rise in total employment of 69,100 over the year ended April, more than two thirds, or 48,700, was in Western Australia.
The implication for job seekers is clear: go west.
The rest of Thursday's jobs report by the bureau had economists scratching their heads. The national jobless rate - the number of jobless people expressed as a percentage of the total number of people either employed or looking for work - fell from 5.2 per cent to 4.9 per cent, an unusually large fall in one month.
This was particularly odd given total employment rose only 15,500, pretty much in line with the historic average monthly gain over the past 20 years.
The super-sized fall in the jobless rate was made possible by a fall in the participation rate - the proportion of the working-aged civilian population either in work or looking for it - from 65.3 per cent to 65.2 per cent.
Why is the participation rate falling if there are more jobs on offer? A very good question and one to which economists have, largely, drawn a blank.
Perhaps all the gloomy talk on the economy convinced some job seekers to give up the search. Perhaps the location and skills of the jobs on offer - mostly in Western Australia - simply did not match the needs of job seekers, so they gave up.
Another possibility is there was a disproportionate increase in the size of the working-aged population. But with the retirement of the first baby boomers, and fewer Generation Ys to replace them, you would expect the working-aged population to be shrinking, or at least slowing in growth.
Over the past year, the participation rate has fallen from 65.6 per cent, reversing a longer-term trend of higher participation. Economists estimate that had participation not fallen, unemployment would be in the mid to high 5 per cent range.
Amid all the confusion, it's important to remember how these numbers are collected. Every month, the bureau phones 30,000 houses to ask them about the working lives of people in that home. This usually ends up covering about 50,000 to 60,000 people, about 0.3 per cent of the population.
To ensure the sample surveyed reflects the broader characteristics of the population, the bureau then assigns each person's responses a relative weighting, depending on whether it needs more or less of that sort of person in its survey. If the bureau happened to get vastly more female respondents than men, it would downgrade the relative weighting of the answers of those women relative to the answers of men's.
This becomes problematic. In doing its weighting, the bureau must make certain assumptions about the population growth, interstate migration and overseas migration. But surveys of population and migration are conducted less regularly than the monthly labour force survey.
Turns out, in an admission by the bureau in this month's release, the sharp fall in student migration during 2010 in response to changed migration laws caught it by surprise. This led it to overestimate the size of job gains in 2010 and, correspondingly, underestimate job gains in 2011.
Economists are now reassessing the strength of the job market, with some arguing that had the true figures been known, the Reserve Bank would have been less inclined to deliver its 50 basis-point cut to the cash rate last week. An interest rate rise next month is now believed to be less likely.
Perhaps they should have asked a taxi driver.