Click here for this post in PDF: Debtwatch members; CfESI members
Click here for the data in this post: Debtwatch members; CfESI members
The most recent ABS data implies that unemployment is falling, and that those who saw the economy as either stable or booming were right. Both the seasonally adjusted and the trend estimates fell, and the seasonally adjusted unemployment is now 0.2% below the peak it reached in August 2011 of 5.3%. This looks good—though not as good as conventional Neoclassical forecasters were expecting a year ago: in the 2011-12 budget, the Treasury, using its TRYM model, predicted a smooth fall in unemployment from 5% in June 2010 to 4.75% in June 2011 and 4.5% in June 2012.
Figure 1: ABS Unemployment Data and Treasury Forecasts
Clearly the Treasury didn’t expect the rise in unemployment that occurred in mid-2011. But the last five months of seasonally adjusted ABS data appear to imply that this deterioration was an aberration, and the expected recovery is on its way once more.
Figure 2: Treasury forecasts (& projections of a return to equilibrium) in the 2011-12 MYEFO
Or is it? As I noted in my most recent post, the definitions of employment and unemployment are now seriously compromised. The formal definition of unemployment used by the ABS is:
Persons aged 15 years and over who were not employed during the reference week, and:
The definition of “actively looked”:
Includes writing, telephoning or applying to an employer for work; answering an advertisement for a job; checking noticeboards; being registered with Centrelink as a jobseeker; checking or registering with any other employment agency; advertising or tendering for work; and contacting friends or relatives.
The definition of employed is:
All persons aged 15 years and over who, during the reference week:
On these definitions, people who are discouraged by the job-seeking process—so that they haven’t applied for a job in the previous 4 weeks—are not unemployed. If they have worked for one hour or more in the previous fortnight, they will be classified as employed; if not, they will be “Not in the Labour Force”.
As noted in the previous post, these definitions disguise the real level of unemployment, and they inspired Roy Morgan to conduct its own survey with a rather more straightforward definition:
The Roy Morgan survey, in contrast, defines any respondent who is not employed full or part-time and who is looking for paid employment as being unemployed. ” (Roy Morgan, September 2001)
When Roy Morgan reports a trend in unemployment, this can be taken seriously—with caveats about the smaller size of the Roy Morgan sample (4,500 versus 30,000 for the ABS), and the fact that the data is not seasonally adjusted. When the ABS reports a trend, it could be a trend, or it could be an artefact of its definitions.
Curiously, the ABS itself implied in today’s statement that its unemployment data should be taken with a grain of salt. They recommended instead using the ratio of employment to population
A different method of analysis that removes the effect of population growth is to compare average employment to population ratios for each year… This analysis provides an alternative comparison of employment between years, as the influence of change in the underlying population in each year is removed.
The ABS’s definition of employment is not above reproach—including as employed people who worked a ludicrous 1 hour in a fortnight (and unpaid at that, if in a family business or on a farm)—but it lacks the additional distortion of its unemployment definition, which classifies the discouraged unemployed as “not in the workforce”. Curiously, the trend to falling unemployment in the last 5 months of ABS data isn’t mirrored in the employment to population ratio: though it blipped up in the most recent month, it has been steadily falling since the beginning of 2011.
Figure 3 illustrates this by graphing the employment to population ratio and unemployment together, and inverting the unemployment data (putting low unemployment at the top and high at the bottom).
Figure 3
A smoothed plot makes the trends in both series more obvious: the employment to population ratio is still trending down—implying rising unemployment—while the recorded unemployment rate is falling.
Figure 4: The data in Figure 3 smoothed
Roy Morgan’s measure of the unemployment rate, however, is clearly increasing.
Figure 5
Roy Morgan’s estimate that 10.3% of the workforce is unemployed is now more than twice the ABS’s estimate of 5.1%, and the gap between the two is the largest it has ever been.
Figure 6
Part of this huge gap in the January 2012 figures is undoubtedly due to the fact that the Roy Morgan data is not seasonally adjusted, and January necessarily involves a huge boost to actual unemployment as school leavers enter the workforce. But the trend in the gap can’t be explained away, and that gap is now the biggest it has ever been.
Figure 7
So it’s too early to declare that unemployment is falling—especially when that call is based upon data with a dodgy definition.
Oh, and that Treasury forecast shown in Figure 2, of rising unemployment from mid-2012—with it returning to 5% from the forecast low of 4.5%? That’s not a forecast: that’s an assumption. Built in to the TRYM model—and every other neoclassical macroeconomic model on the planet—is the assumption that the economy will return to a long run equilibrium rate of growth after any short term “shock”. The Treasury happened to assume that this long run equilibrium includes an unemployment rate of 5%.
This assumption is a classic example of the adage that “to assume makes an ASS out of yoU and ME”. Neoclassical economists have been getting away with this sort of behaviour for decades, because the public didn’t challenge them before the Global Financial Crisis when the economy seemed to be doing well. I hope now that, after the crisis, the public will be as critical of such assumptions as Roy Morgan has been of the ABS’s dicky definition of unemployment.
