Seeking Alpha
Profile| Send Message|
( followers)  

Thursday's release from ADP showed a net decrease of 33K jobs from the US economy. While I have greater confidence that the ADP sample is more accurate than the same numbers from the BLS that will be released on Friday, this number is still too close to zero to have any level of confidence that there was an actual net increase or loss of total jobs. In addition, each month's release is revised at the next month's announcement. With a number this close to zero, the revision could be randomly negative or positive. Which means that the market's strong reaction to the data may be without basis.

Advanced Data Processing [ADP], the publisher of this report, is one of the largest outsourced payroll processing firms in the US with more than half a million clients worldwide. ADP analyzes their payroll data each month and releases a statistical estimate of the net loss or gain of jobs in the US and within specific employment categories. The Bureau of Labor Statistics releases an identical report based on employer submitted unemployment insurance information. ADP releases their report on the Wednesday before the first Friday of the month. The BLS releases their report (commonly referred to as the Non-Farms Payroll Report) on the first Friday of the month. Because both reports are statistical samples they will never be identical. The fact that they are both statistical samples is important to understand as we seek to use the information in our investing.

A number that is supposed to reflect a large population based on a statistical sample will have confidence intervals associated with it. That means that because the statistician does not know what the actual number of jobs lost is they pick a range, within which, they are confident the real number exists. The 90% confidence interval for the labor report from the BLS is plus or minus 430K jobs. That means that statisticians are 90% confident that the real number of jobs added or lost from the economy is somewhere within a range of 860K jobs. This puts the -33K jobs number today into perspective. It is well within the realm of possibility that there were 50K jobs added or even 100K lost. The statisticians simply do not know what the real number was and the narrower the range is the less confident you can be.

Despite the fact that many experienced traders are well aware of the reliability issues in the US labor reports we still saw a robust negative reaction in the US equity market to the numbers, which bled over into the forex and bond market. The report was blamed for much of the price breakout on the S&P 500 but can we expect this trend to continue based on this data?

Source: The Market's Reaction to Unemployment Data