It’s taken just three months of the new, imprecise retail trade trend series of figures to force the Australian Bureau of Statistics to reverse course and go back to the old, more detailed series with a larger sample. The backflip was announced yesterday in a statement from the ABS.
Chief statistician Brian Pink said in that statement that the decision followed global financial market developments in recent months that had resulted in closer scrutiny of economic data.
“In addition, key macroeconomic statistics users had indicated that more robust monthly retail trade data are their top priority at this time for improved economic statistics,” the statement said.
Could that have been the chaps at Martin Place who were becoming concerned at the difficulty of working out what was going on in retailing?
Get Crikey FREE to your inbox every weekday morning with the Crikey Worm.
After all, the trend series has been showing an increase of 0.2% in retail sales since July and yet a low key, seasonally-adjusted figure that is buried in the new figures each month (because it’s not really as accurate as the pre-July seasonally adjusted figures) was quickly picked up and used by economists and commentators.
This week the September retail sales trends seasonally adjusted figure showed a very sharp fall of 1.1%, something at odds with the broader trend.
The seasonally adjusted fall was more in keeping with the slowdown in same store sales being reported by retailers like Harvey Norman and Clive Peeters: their same store sales fell in late September and into October.
“Whilst month-on-month changes in retail sales are difficult to measure reliably, even with the best of survey designs and robust samples, reinstating the full monthly sample will reduce some of that uncertainty,” Mr Pink said in the statement yesterday
Earlier this year the ABS reduced its survey sample size by two thirds after a $22 million cut to its budget.
Economists had criticised the decision at the time, saying the market needed to have reliable indicators of consumer spending to evaluate overall growth in the economy.
The ABS warned on Wednesday that it will face extra costs by reinstating the sample size, but does not propose to make offsetting cuts.
Results from the reinstated sample will be available in early 2009, which means we will get up date figures for the important Christmas period. Quarterly retail trade figures with a solid seasonally adjusted figure will be released by the ABS on November 17.
Now for the other re-worked statistical series, today’s Labour force figures, which also suffered cutbacks. Because the sample for this was shrunk to save money, there’s some doubt as to whether it has accurately picked up changes in the employment market since July when the new approach started.
The ANZ Jobs ads series, which is a solid predictor of future changes in employment, has been showing a worsening in the state of the jobs market. The first hint we had from the ABS Labour Force series was in the September report, but even that was at odds with the trend shown in the ANZ series.