by Calculated Risk on 10/24/2012 06:50:00 PM

ADP announced some changes to their monthly employment report today (ht Rob). Here is the press release, and here is the website with FAQs, methodology and more:

The newly expanded ADP National Employment Report will be issued each month by the ADP Research InstituteSM, a specialized group within ADP that provides insights around employment trends and workforce strategy. The first enhanced monthly report issued in collaboration with Moody’s Analytics will be released on November 1, and will report private payroll changes for the month of October 2012.
...
This new collaboration allows the ADP National Employment Report to increase the number of industry categories reported and expands the number of business sizes reported each month. Other key enhancements of the report include the development of a new methodology to further align with the final, revised U.S. Bureau of Labor Statistics (BLS) numbers. A look back at historical data from 2001 to present using the new methodology shows a very strong correlation (96%) with the revised BLS numbers. In addition, the overall sample size used to create the report has been increased from 344,000 U.S. companies to 406,000, and from 21 million employees to 23 million; which accounts for more than 20% of all U.S. private sector employees. Originally launched in 2006, the ADP National Employment Report is a derived from actual payroll data from an anonymous subset of ADP’s clients in the U.S.

Basically ADP/Moody's uses the ADP payroll data and attempts to predict the BLS report of private sector payroll jobs added or lost each month. They use matched pairs (companies that report for both the current month and the previous month), and adjust the data by category to align with the BLS data (final data after revisions).   It is a fairly complicated process.

However the ADP employment report is still a black box and we will not be able to judge the "improvements" for a few years. I think it would be much better for analysts if ADP just reported actual payroll gains or losses by industry (perhaps in percentage terms). That would be a useful and independent measure of the labor market.

The first "enhanced" report will be released next week on Nov 1st.

Earlier:
• New Home Sales at 389,000 SAAR in September
• New Home Sales and Distressing Gap
• New Home Sales graphs