- Telephone
- Email
Just after Christmas The Cairns Post ran an article highlighting the differences between our own Conus Trend and that calculated by Rick Carr at Herron Todd White, when we both came up with some significantly different numbers for the Nov data (see my post on the article here). Today sees the release of HTW’s Cairns Watch report for January which includes the Dec Cairns region unemployment data (you can download the report here, and read our own post on the Dec data here).
What becomes clear is that the confusion caused by the problems the ABS have been having with their seasonal adjustment of employment data is making it very hard for the two of us to agree on what is going on in Cairns. Rick has the Trend unemployment rate in Cairns in Dec at 9.0% while our own Conus Trend data shows a rate of 7.6%. So who’s right? Well, frankly we won’t be able to answer that question until we see how the next few months of data pans out. If the move up that we saw in the original data in Dec is repeated in Jan and Feb then we can expect to see out own Conus Trend move higher (and the past readings also revised up). If, on the other hand, we see the original ABS data moving back down in Jan and Feb we would expect to see Rick’s Trend series fall sharply (with downward revisions to previous months). There is no simple answer, although obviously we prefer our own series!
The two graphs below (the second is taken from HTW’s Cairns Watch report) might help readers to get an idea of the difficulty that the two models are having, and the way in which both are trying to solve the problem. Trying to fit a trend line to such a volatile series as the original ABS data is always going to be a challenge! What is also clear that is, until the middle of 2014, the two series were actually tracking very closely to each other.
We are proud to have CBC Staff Selection as supporters of the Conus/CBC Staff Selection Trend Employment and Jobseeker data series
February 2nd, 2015 at 7:49 pm
Do you know what model/filter Rick is using to generate his trend series? Eyeballing it, it doesn’t appear to be generated from any I’ve seen.
February 2nd, 2015 at 8:50 pm
No I don’t Gene. Rick’s Cairns Watch report also doesn’t explicitly note revisions, although looking at graphs from previous reports it is clear that they are made. My own Conus Trend is created using the X-12 ARIMA model.