- Created on Tuesday, 02 August 2011 17:02
- Written by Jared Levy, Editor, Option Strategies Weekly
- Hits: 2411
If you are a long-time reader of Smart Investing Daily, you know my predictions on the monthly Bureau of Labor Statistics (BLS) jobs report are pretty accurate.
The Bureau of Labor Statistics jobs report is released on the first Friday of every month, and sets the benchmark for employment here in the US.
It is the most important piece of monthly economic data, and this Friday we'll get a good look at where we stand.
ADP, the largest payroll provider in the U.S., releases its report two days before the Bureau of Labor Statistics. Investors watch this report for a "sneak peek" at what the jobs report will look like on Friday.
Last month, ADP's employment report surprised a lot of analysts, including me: 157,000 new positions were created. This was tremendous job growth. The report sent the markets skyrocketing and tens of thousands of investors bought into the rally.
The S&P made a two-month high that day...
But the Bureau of Labor Statistics' jobs number was a rude awakening... the worst job growth in eight months. The ADP number was completely wrong! After the report, the S&P 500 lost 5% of its value.
How could ADP be so far off?
One of my mentors taught me a lesson about data models a long time ago. He said, "Bad data in, bad data out." The biggest problem that investors face is bad data.
You'd be surprised at how much data is actually collected and how much is merely estimated.
The Bureau of Labor Statistics
The Bureau of Labor Statistics' survey sample includes about 140,000 businesses and government agencies that cover 440,000 worksites. In all, the Bureau of Labor Statistics samples roughly 9.0 million Unemployment Insurance tax accounts.
This figure only equals one-third of all nonfarm payroll employees. The Bureau of Labor Statistics uses those results to estimate total unemployment.
The figure is incomplete in lots of ways.
It only counts people who got paid through the 12th day of the month. If your pay period doesn't include the 12th, you're not counted!
Also, the Bureau of Labor Statistics does NOT include proprietors, the un-incorporated self-employed, volunteer or family workers, farm workers, domestic workers, or members of the military. Non-civilian government workers are also excluded.
The ADP started reporting in 2001. After years of getting a bad rap, ADP revamped its report to more accurately predict the BLS report.
The new ADP National Employment Report is based on anonymous payroll data from about 400,000 U.S. businesses. Around 23 million employees across all private sectors and regions are included in this report.
This means that the reach of ADP's report is broader than the Bureau of Labor Statistics.
So why was the ADP report so different from the Bureau of Labor Statistics report?
ADP estimates payrolls that don't include the 12th of the month and they use fancy math like regression models to guess (they call interpolate) what that payroll might have been.
Remember, the Bureau of Labor Statistics doesn't include these payroll figures.
ADP also takes a different sample of Americans, which might affect the results
Also keep in mind that ADP's data is ALWAYS seasonally adjusted. (The BLS offers both adjusted and non-adjusted.) I think there are some of the reasons why ADP's estimates were off.
Which Job Report Should You Trust?
Since ADP enhanced its formula, the report has typically been LESS optimistic than in the past. This is why traders went hog wild on last month's ADP report.
Erik Dellith ran some correlation numbers on the two that were eye-opening.
Over the long term, the ADP report is 92% accurate, which is really good. If I knew that an indicator was correct 92% of the time, I would feel pretty darn confident using it.
The problem is that the ADP report seems to lose accuracy when there is a recovery with little or no real jobs being added to the economy. This is exactly the climate we're in right now.
In times like these, accuracy drops to about 70%, which makes me uncomfortable when I am investing my hard-earned money. To put it another way, it's like crossing a bridge that had a 30% chance of collapse.
But I say both reports are flawed. But because the Bureau of Labor Statistics report is the standard that we all rely on, the ADP report is just a "predictor."
Not Good Enough
When most people look at these numbers, they might tell you that ADP is a "good" forecaster of the Bureau of Labor Statistics report. For me, as an option trader especially, I want to know just how "good" that something is.
The ADP number has been accurate about 70% of the time over the past two years. You may want to think twice about betting big on its forecast and getting caught on the wrong side of the data.
Because the ADP report was so strong last month, we may see a drop in Wednesday's jobs report. I would also look for weakness in the Bureau of Labor Statistics numbers again on Friday.
When I make an investment, I calculate "odds" of success to a T. In my trading service, WaveStrength Options Weekly, I use options spreads to increase probability of success and give my subscribers an edge.
We can achieve an 85% chance of success with this strategy. That's much higher than the ADP's jobs report.
We need an "edge" in every investment that we make. We get it through better information, a better strategy or a combination of both. In every trade you make, be sure you know your odds and have your own edge.
Publisher's Note: Jared is a master of numbers. His expertise goes far beyond deciphering the truth behind the government's numbers. For his subscribers, it means profits. Like Jared says, his options trading strategy has an 85% success rate.
His readers are booking gains of 20%, 40% or even 50% almost every week of the year. To see how they do it, follow the link.
Article brought to you by Taipan Publishing Group. Additional valuable content can be syndicated via our RSS feed. Republish without charge. Required: Author attribution, links back to original content or www.taipanpublishinggroup.com.
Other Related Sources: