Last year, the monthly jobs report was undermined by the Australian Bureau of Statistics’ attempt to change the way it collects and measures the responses from people in its monthly panel. The ABS spent months trying to fix it, eventually succeeding, and the monthly reports on employment settled down, and analysts and the markets started accepting them. But, in May labour force figures, new problems seem to have emerged in Western Australia.

The jobs data report said:

“In compiling this month’s estimates the ABS identified that the growth in employment for Western Australia, in original terms, was unparalleled compared to previous May results. Investigations showed that a significant contributor to this was change in response patterns from April to May 2015. As a result, the ABS has removed this impact from the seasonally adjusted estimates for Western Australia and subsequently, Australia. As the ABS is unable to remove this impact from the original estimates, care should be taken in comparing the original and seasonally adjusted estimates, particularly for Western Australia.

“The ABS has not made any changes to its Labour Force Survey methods or processes that could have caused the increase. Estimates derived from sample surveys will occasionally be subject to significant statistical volatility. The ABS is continuing to investigate sampling and estimation options to reduce the volatility in surveys like the Labour Force.”

And why is this important? The jobs report for May was much, much better than expected and at variance with the flow of other economic data, except for the dwelling construction sector. The jobless rate (seasonally adjusted) fell to 6% (the federal Treasury and RBA are forecasting it to rise to a peak of 6.5%) as 42,000 new jobs were reported, 22,000 people dropped out of the unemployment pool (all looking for full-time work), the participation rate was unchanged from 64.7% and the hours worked rose just 0.1% in the month from May.

You could call it too good to be true — but there is just enough credible information in the jobs report to give it credence. But those measurement problems in WA are a big, big worry for the most important monthly economic figure we have.