Non Farm Payrolls: The Devil Is In the Adjustments
by ilene - September 6th, 2010 5:15 am
Non Farm Payrolls: The Devil Is In the Adjustments
Courtesy of JESSE’S CAFÉ AMÉRICAIN
When the US government announced a ‘better than expected’ headline growth number in its non farm payrolls report for August, a loss of ‘only’ 54,000 jobs versus a forecasted loss of 120,000 jobs, people had to wonder, ‘How do they do it? We do not see any of this growth and recovery in our day to day activity.’
Here’s one way that those reporting the numbers can ‘tinker’ with them to produce the desired results.
As you may recall, there is often a very large difference between the raw, unadjusted payroll number and the adjusted number. Seasonality plays the largest role, although there can occasionally be special circumstances. Since this is designed to be a simple example I am going to lump all the various adjustments that could be and call them the ‘seasonality factor’ since it is most usual and signficant.
Here is a chart showing the unadjusted and the adjusted numbers. As you can see, a seasonal adjustment can legitimately normalize the numbers for the use of planners and forecasters. This is a common function in businesses affected by seasonal changes. Year over year growth rates, rather than linear, comparisons, can also serve a similar function.
Quite a variance in numbers that are very large.
Since it probably is in the back of your mind, let’s address the infamous "Birth Deal Model" now, which I have advised may not be such a significant factor as you might imagine. This is an ‘estimate’ of new jobs created by small businesses. A comparison of the last few years demonstrates rather easily that this number is what is called ‘a plug.’
How can the growth of jobs from small business not been significantly impacted by one of the greatest financial collapses in modern economic history?
Certainly the Birth Death model offers room for statistical mischief. It is important to remember that it is added to the RAW number before seasonal adjustment, and that number has huge variances. So the effect of Birth Death is mitigated by the adjustment for seasonality. If it were added to the Seasonal number from which ‘headline growth’ is derived it would be a huge factor. But it is not the case, although the timing of the significant annual adjustments and additions is highly cynical, and supportive of number inflation.…