- Telephone
- Email
Readers of the Cairns Post this weekend might be justifiably confused about what is actually happening with employment in the region. Following the release by the ABS of regional labour force data on Thursday the Post ran a story on Friday talking about the fact that Cairns had “added 6,100 jobs to its workforce in the year to March” and quoting Cairns MP Michael Healy saying “This is no blip. The trend has been clear for months now that unemployment is tracking down.”
How then does that story tally with our own commentary on Thursday (see here), where we noted employment down 600 over the year, and the follow-up story in the Cairns Post on Saturday which noted that according to our Conus/CBC Staff Selection Trend we had seen an “increase in the Trend unemployment rate over the year..” of 1.4%?
To understand the apparent paradox we need to realise that the official figures to which the Cairns Post and Michael Healy were referring on Friday are the simple average of the previous 12 months of original ABS data.
Each month the ABS conduct a Labour Force Survey to determine how many people are employed and unemployed throughout the nation. To do so they survey a sample of the over-15 year-old civilian population. The survey covers approximately 0.32% of that cohort; in other words about 64,500 people. At the national and state level the ABS then use the numbers straight from the survey (the “original” data) and, by the application of some statistical techniques, create a seasonally adjusted and Trend series which aim to remove some of the volatility inherent in the original data. So, for example, in the latest March data the original employment data had 12,491,700 people employed; the seasonally adjusted series came in at 12,484,100; while the Trend was 12,485,800.
At the level of Statistical Area Level 4 (SA4) Cairns the volatility in the original sample (which could be only about 650 people) can be very large. The ABS therefore recommend, in the absence of anything better, using a simple 12 month average (12ma) of the original data. It is this 12ma that was being referred to in the Cairns Post‘s Friday story.
Some years ago we became frustrated with the way in which this simple measure lags real changes on the ground and worked to do what the ABS have lacked resources to do; create a Trend series for the Cairns labour force (this was subsequently expanded to include all QLD SA4s and now also includes NSW). In doing so we utilise a model very similar to that used by the ABS at the national and state level to create the Conus/CBC Staff Selection Employment Trend series. Using a Trend essentially allows us to “see through” some of the inherent volatility while not losing the effect of real on-the-ground changes (as the 12ma does). The difference in the way in which the two measures pick up changes is obvious,the chart below makes this clear.
The dotted red line is the original ABS data (note the volatility) while the green line is the simple 12ma. The blue line is the Conus/CBC Staff Selection Trend and. as we can see, this tends to identify changes in direction far earlier than the 12 month average. Indeed on average the Trend spots the turning points some 6-7 months earlier. Note that the 12ma appears to have just noted the turnaround that the Trend identified back in Sept 2017.
The “6,100 increase in jobs for the year to March” relates to the simple fact that the average for the 12 months to March 2018 was 6,100 higher than the average of the 12 months to March 2017. Looking at the original data we can see that a large part of the reason for this was that the data for May, June and July 2017 were unusually high compared to unusually low data in May, June and July 2016. You might ask the question “how relevant is that fact to what is actually happening in Cairns in March 2018?”; we would suggest an answer of “very little”. The reality is that the simple 12 month average provides only a very lagged indicator of labour market strength; it certainly is never used at the national or state level where “better” Trend and seasonally adjusted series exist.
There are two other facts that we must acknowledge.
Firstly, Trends by their very nature are subject to revision once we have future month’s data. This means that, were the original employment data for Cairns in April to come in significantly different to 116,000, we would expect to see the March data revised either up or down. For that reason we always caution against reading too much into individual months’ of Trend data and would rather wait for 2 or 3 months to confirm directional shifts. However, as we see in the chart above, this downward move in Cairns employment has now been in place since Sept 2017. It is clear therefore that Michael Healy’s comment that “the trend has been clear for months” is correct, unfortunately in the opposite direction than to which he referred!
Secondly, none of these measures tell us exactly what is happening. The original data is gleaned from a relatively small sample and as such is simply our “best guess” at what’s going on. To get a truly accurate picture would require a Census; and even that would be out-of-date before we started counting! In considering measures of the labour market we have to determine which gives us the most timely and potentially accurate story of reality. The ABS is clear that, where available, we should be using Trend series to consider employment data at the national and state level. The Conus/CBC Staff Selection Employment Trend allows us to do the same at the SA4 level and is, in our opinion, a far better indicator of reality that the “official” 12 month average.
We are proud to have CBC Staff Selection as supporters of the Conus/CBC Staff Selection Trend Employment and Jobseeker data series