On Thursday the Australian Bureau of Statistics (ABS) will release Australia’s jobs report for June.
To say that mistrust in the ABS’ seasonally adjusted figures is elevated is an understatement.
Years of wild monthly swings, creating uncertainty rather than certainty when it comes to developments in the labour market, has left the release only useful to day-traders looking to capitalise on short-term market movements.
It’s got to the point where the former head of the ABS described the report as not worth the paper it’s written on. Hardly a ringing endorsement.
Well, the wild swings may be about to get that little bit wilder in the coming months.
Sample rotation effects, something that contributed to the enormous employment gains seen in October and November last year, are about to exit the data series, suggesting to the National Australia Bank (NAB) that a big downside miss may be on the cards in June.
Very strong rises in employment and significant declines in the unemployment rate were recorded in October and November 2015. In those months, the incoming sample cohort accounted for around 60% of the jobs growth, despite comprising only one-eighth of the total sample (those cohorts had a higher employment intensity than the existing sample).
The ABS uses eight rotation groups to conduct the labour force survey. As a new month arrives, a new survey group enters the sample pool, with the group that has been in the survey the longest dropping out of the data series.
As the NAB points out, around 60% of the jobs created in October and November last year came from just 12.5% of the survey sample, due to those groups entering the series having a higher rate of employment than those which they replaced.
The chart below, supplied by the NAB, shows the contribution to overall employment change by the incoming survey rotation each month, going back to February 2015. Those contributions to employment change by the incoming sample group are shown in red.
Now, eight months on, those groups that contributed to hefty employment gains late last year are about to exit the sample rotation, indicating to the NAB that a large drop in employment could eventuate.
Those sub-samples begin rotating out from this week’s June Labour Market report, setting up the possibility for some weaker employment and higher unemployment reads over the next two months. NAB estimates the sample rotation effect as potentially subtracting as much as 30k jobs a month over the next two months and that the unemployment rate could tick-up by 0.15% points. While the effect is far from certain (and dependent on job creation in the existing sample, the characteristics of the incoming sample, along with any ad-hoc smoothing the Statistician may apply), it is significant enough for NAB to forecast a well-below market outcome for June employment and will complicate assessment of the underlying trend in the labour market.
As a result of the survey rotation, the NAB is forecasting a decline in employment of 17,000 in June, well below the median consensus forecast offered to Bloomberg which looks for an increase of 10,000.
It also expects that the unemployment rate will tick up to 5.8% from 5.7%, in line with median economist expectations.
While such a headline result would normally be perceived as negative, the NAB isn’t concerned, suggesting that other complementary labour market data is trending in a positive direction at present.
“We continue to assess the labour market as broadly positive, with the NAB survey’s employment index pointing to employment growth of 15-18k a month, and SEEK job ads have also picked up a little in recent months,” its says.
That final statement sums up the trust in the ABS survey in a nutshell. Given doubts about its accuracy, economists and markets are now turning to other alternate surveys to gauge what is really happening in the Australian labour market at present.