How unemployment is really calculated...

R

retaildaytrader

The government report is just a sampling of about 110,000 people...some people are surveyed in person while others are surveyed over the phone. In any event, it is basically a statistical guess relying on a sampling by asking questions of people in one on one interviews.

http://www.bls.gov/cps/cps_htgm.htm

The ADP report is derived from a sampling size of 22 million people using hard data from the payroll system (versus manual interviews).

http://www.adpemploymentreport.com/pdf/FINAL_Report_March_10.pdf

The key difference is the ADP employment data is all non-government employees.

This month the ADP employment data suggested a decrease of 23000 jobs in the economy where the government's data suggested an increase by 168000 jobs.

So you decide which report is correct and which is not correct...I believe Wall Street uses the government's labor report, but my sincere belief is the ADP report is by far a more accurate picture of the US economy.

Its funny because the government and ADP could simply partner together and save millions of dollars by creating one single report utilizing both their computerized payroll records. Instead of using the census bureau to take manual surveys, they could do it together using hard data from payroll records. The millions saved could then be used for such things as scholarships and small business lending...
 
You didn't read the description of the ADP methodology closely enough. ADP report is also based on a sample, rather than the totality of private payroll data. The sample is somewhat larger, to be fair (currently at arnd 400,000), but it's also a "statistical guess".
 
Thats 400,000 US Business clients not people...the number of people is 22 million in the sample.

"The ADP National Employment Report, sponsored by ADP®, was developed and is maintained by Macroeconomic Advisers, LLC. It is a measure of employment derived from an anonymous
subset of roughly 500,000 U.S. business clients. During the twelve month period through December 2009, this subset averaged over 360,000 U.S. business clients and over 22 million
U.S. employees working in all private industrial sectors."
 
I could be wrong, but this is the passage I am referring to (in the methodology document):
"The resulting sample of payrolls currently is approximately 400,000, covering roughly 23 million employees in all major private industries and regions."

I took it to mean that it's 400,000 individual samples that are taken from, or "cover", the 23 million total that is being subjected to the survey. I could be wrong, obviously.

Besides, I am not entirely sure what your point might be. The NFP release includes govt employment, as well as private, so that's one big difference. As to the rest, according to ADP/Macroeconomic Advisors, ADP numbers are very closely correlated to the BLS numbers, so, in general, they seem to agree.
 
You must log in or register to reply here.
Back
Top