Monday, March 24, 2014

Brotherhood of St Laurence jumps the shark on youth unemployment

The Brotherhood of St Laurence have in recent months been campaigning on the worthy cause of youth unemployment. The most recent update from the Brotherhood is now being widely reported with some sensational growth forecasts: No jobs or attitude? Why are so many youths unemployed in Cairns?
Youth unemployment in Cairns is now the highest in the country and is reaching a crisis point.
Figures released by the Brotherhood of St Laurence yesterday show an 88 per cent rise in the jobless rate over the past two years and youth unemployment in the city at 21.6 per cent as of last month.
That figure is set to more than double by 2016 to 46 per cent if action is not taken.
The numbers here are derived from the ABS labour force survey. We know how volatile this is for the entire region, or even states, with frequent cautions on the sample size and even more caution is required for smaller subsectors such as youth employment. Conus has posted in recent months on the problems and issues here with the data and also applied a trend analysis, although still cautions on the reliability of this with some unrealistic anomalies apparent in the underlying data:

Cairns Youth Unemployment

Brotherhood of St Laurence and Cairns Youth Unemployment ...again
Youth unemployment in Cairns is clearly a problem but the Brotherhood's reporting, and seeming lack of understanding of some of the statistical difficulties, doesn't appear to be a particularly helpful way of progressing the issue.

I think that is too kind. I would actually call it beyond naïve to the extent of dishonest and appears to have been manufactured with deliberate intent for sensational media coverage.  The particular issue is the way they have taken unreliable regional data to extrapolate sensational "forecasts" of extreme levels of youth unemployment in two years.

The emphasis in the media release from the Brotherhood is on growth percentages and their extrapolated numbers for media consumption:
If it continues to rise at the current rate, the level of youth unemployment will hit 46 per cent in parts of Australia in 2016, the revealing analysis of the new data found.
Somehow we then move subsequently in the media release from this already stretched hypothetical to an analytical "forecast":

At these rates of increase, the new analysis forecasts youth unemployment by 2016 will hit:  46.0 per cent in Cairns.
This methodology is garbage in - garbage out squared! Had they extrapolated like that for a few more years they could have had a youth unemployment rate beyond 100% and really sent Nick Dalton into orbit?

Detail of their "analysis" is actually somewhat limited beyond the media releases so it isn't even clear how they have reached some of these numbers from what they have provided. I wonder if they would like to include underlying assumptions or implications on broader unemployment rates for the period?

I seem to recall a previous post on another worthy charity misrepresenting homeless statistics.

Update: Comments and analysis on youth unemployment from Jeff Borland at Melbourne Uni.

1 comment:

  1. Yeah, using the detailed regional data in this way to "forecast" is statistically just plain wrong. Looking at the detailed ABS data you can see what they have done. For the 12 months to February 2014, the average youth UE rate was 21.6% and for the 12 months to February 2012 it was 11.5%. Extend that trend out to 2016 and you get the 46% number. Extend it another 2 or 3 years and you go over 100%. This is obviously nonsense and complete rubbish. What they did not tell you is that if you look at Qld Outback region, the UE rates have gone down and was 26.6% in the 12 months to February 2012 and 19.8% in the 12 months to February 2014. "Forecast" that trend and UE in the Qld Outback will quite quickly fall away to much less than the national average! the same goes for a lot of other regions. It just goes to show how highly variable the data are at he regional level.