A sharp jump in full time jobs underlines the case for ignoring the seasonally adjusted employment data and watching the trend series in the February report from the Australian Bureau of Statistics yesterday.
But, the trend series showed a much larger than expected rise in new jobs, adding to the feeling that the economy is continuing to strengthen.
However, there could be another reason for this strength. We explain that below.
The seasonally adjusted series showed a huge 47,000 jump in new jobs – with all of that coming from an estimated 80,500 jump in full time employment, which more than offset a 33,300 drop in part time employment. The news helped push the value of the Aussie dollar back towards 90.50 USc – up half a cent, after a solid rise overnight.
That puts the Aussie a little bit further away from the rising Kiwi dollar after that country’s central bank lifted interest rates by 0.25% to 2.75% earlier today.
Further raising questions was the similarity between this February’s figures and last February’s when there was a big rise in new jobs, which in that case was driven by a surge in part time work.
That was one of the largest increases in full time employment seen for years, and is similar to the 71,000 plus jump reported in February, 2013 – that time it was a huge jump in part time jobs (almost 54,000 which added to a near 18,000 rise in full time employment). But then the trend series showed a solid rise of more than 30,000 new jobs in the month.
Last month it was a huge rise in full time employment – coming after January had seen a fall of around 7,000 in full time jobs, partially offset by a small rise in part time employment (for a small fall in job numbers overall). But January’s fall of 3,700 jobs was revised to 18,000 new jobs, seasonally adjusted, in yesterday’s report.
Normally that would be a solid report for any month of the year and means around 66,000 new jobs have been created so far this year – which is a much stronger start than the 5 decade high unemployment rate would indicate. But it also echoes the solid start to job creation in 2013 which then petered out as the year went on.
Further adding to the unworldly nature of the data was a seasonally adjusted fall in hours worked in February, which seems distinctly odd given there was such a huge increase in full time jobs. The participation rate rose 0.2 of a percentage point to 64.8%, which is a positive move seeing it’s vital to get as many people involved in the workforce as we can.
The new full time jobs works out at around 4,000 a day, while the lost part time work was around 1,300 a day – that made for a very busy labour market last month. The ANZ job ads series saw a 5% plus rise in February, but the accuracy of that month’s survey was clouded by what the bank said were problems with the seasonal adjustment process.
The ABS said in in yesterday’s release that the end result was "Australia’s seasonally adjusted unemployment rate increased by 0.1 percentage points but the rounded estimate remained at 6.0% in February”.
In other words, the unemployment rate almost rose to a new decade high of 6.1%. "The increase in total employment was driven by increases in male and female full-time employment and male part-time employment,” the ABS said.
The number of people unemployed increased by 9,800 people to 742,200 in February.
That’s probably due to more people rejoining the workforce to look for jobs – an outcome perhaps of some positive news in the month – but unexpected given the bad publicity around jobs and Qantas’ problems.
The ABS monthly seasonally adjusted aggregate hours worked series decreased in February, down 14.0 million hours to 1,608.9 million hours. That was down on the 20 million or so extra hours worked in January, when employment fell 3,700 (now a gain of 18,000).
Looking at the trend data, there was a still solid 36,000 jobs created last month, while unemployment rose by around 16,000, the participation rate remained steady on 64.7% and the number of hours worked rose by just over two million.
But that said, there’s a big question mark or three associated with the data, which in turn suggest another reason for the big increase.
Every month the ABS changes one eighth of the households in its survey sample.
This new sample can have a skewing impact (as it seems it did in February 2013 with the much higher than expected new jobs numbers).
Because of the vagaries of random variation, the new samples have different characteristics – more employed people or fewer retirees or stay-at-home parents, or even young people, than the one they replace or the existing groups in the survey.
These change over the year. Overlay this with the seasonal adjustment factors and you can get some hard to believe results.
So the reported levels of unemployment, employment, participation, hours workers (and full and part time work) can change with each change of group through the year.
In the new February group, there was a higher proportion of people in the work force – either employed or unemployed (and looking for work).
"This incoming rotation group contributed, in original terms, 37 per cent of the increase in total employment and 29 per cent of the decrease in persons not in the labour force in February 2014," the ABS said.
If you look at the original figures referred to by the ABS in yesterday’s data – that is, not seasonally adjusted, or trend – they showed employment rose by 139,500.
And around 37%, or about 52,000, came from the change of group with the higher proportions of people in the workforce or looking for work.
Apply that 52,000 figure to the seasonally adjusted figures and you end up with a small marginal change, or no change at all – and you also wipe out most of the gains on the trend series, and probably the January revision.
Remember the jobs data show estimated, not hard and fast figures.
