What's really going on with Queensland's employment numbers?
The October ABS labour force figures showed Queensland’s unemployment rate to be the highest in Australia – substantially above the national average.
I recently heard some pundits and commentators suggesting that the October ABS estimates showing a decrease in employment growth in trend terms is innocuous and rather unremarkable since employment figures go up and down.
And, of course, they do “go up and down”.
However, CCIQ does hold concerns about the October estimate for several reasons.
First, the decrease of about 550 in the total employed persons trend estimates represents the 13th percentile; and a negative growth number has not been recorded since September 2016.
The second is most important.
Since January, this year’s trend data tends to indicate that the headier days of jobs growth observed during 2017 are behind Queensland; whilst growth in employed persons was marginally positive up until October, during 2018 it has been weak and well below average.
Given this predicament, I think it’s important to have a closer look at what’s currently going on in the Queensland labour force relative to what has occurred in the past - after all, that is the most objective way to measure relative performance and to also understand today’s numbers in context.
Accordingly, in this article I will put Queensland’s employment growth numbers into historical perspective.
The following graph shows the cycle of trend employment growth, both in persons and percentage changes, from February 1978 to October 2018. We’ll look at the trend numbers, as these tend to smooth out noise present in the original data.
While there is no way of predicting any one particular level of employment growth each month, however, we can use statistical methods to measure recent employment growth statistics relative to historical observations.
First, let’s look at the frequency distribution of the raw data of which there is a reasonably large sample size of 488 observations.
As the graph below shows, the histogram is roughly bell-shaped and symmetrical in nature – and looks reasonably normal.
Now, to smooth the distribution and normalise it further (both the probability density function and the cumulative distribution function), we can use its mean and standard deviation as well as a random number function to generate a larger sample consisting of 2000 observations as follows.
Having now normalised the distribution, we can calculate z-scores to determine the probability of observing one particular employment growth number or less relative to the historical distribution. In other words, one can determine the percentile in which that one growth figure sits relative to the other observations.
By graphing these percentiles with the trend employment growth data, we can now see both the ABS estimates of trend monthly employment growth and the magnitude of this growth in the context of the full sample of the ABS estimates from February 1978 to October 2018.
But let’s not look too far back and instead focus on the period from 2000 onwards.
The graph shows that Queensland experienced some fantastic years of employment growth during the period 2001 to 2009.
Moreover, the following years were rather patchy including a year of relatively high employment growth during 2017.
The latest October number, however, is a different story - showing a decrease of about 550 employed persons in trend terms.
Recall from the beginning of this article, this decrease represents the 13th percentile; and growth in employed persons has not been negative in Queensland since September 2016.
Moreover, the trend data suggest the headier days of jobs growth observed during 2017 may be over as growth has been weak and well below average during 2018