ADP’s February Employment Numbers: The First Release of a Hopefully Improved Report (Feb Report: +57,000)
After some significant “misses” last year when compared to Uncle Sam’s Bureau of Labor Statistics (BLS) reports that come out two days later each month, ADP has retooled its National Employment Report. If how Biz Weak assesses the upgrade (appears to require paid subscription) is accurate, it could prove more valuable than what BLS releases — but it will take at least a year, and probably two, before we’ll know that:
A Makeover For A Key Jobs Report
An overhauled version of Automatic Data Processing Inc.’s (ADP) National Employment Report is due on Mar. 7. The changes should get monthly results a little closer to the Labor Dept.’s initial monthly payroll figure. But the report’s more valuable contribution may be a better picture of the labor market in real time.
….. The ADP report now looks at nearly 400,000 businesses with payrolls totaling close to 23 million workers. That’s larger than the initial pool used by the Labor Dept. Plus, the data are now being collected weekly, vs. monthly, and a more advanced seasonal adjustment is in place.
The two measures will still diverge from time to time. When that occurs, early evidence shows the ADP numbers come closer to the government’s final annual employment revisions, released in early February, than the initial Labor Dept. jobs data. One reason is that the ADP report includes businesses not in the government’s initial survey. That makes the report “particularly powerful if you are trying to understand what truly happened to employment, as it will eventually be reported by the [government],” says Macroeconomic Advisers Chairman Joel Prakken.
It will take time to convince the financial markets, but if the revised ADP report proves to be a reliable real-time estimate of true labor market conditions, it will end up being quite valuable.
The final revisions referred to in bold are no small matter. The one just completed a month ago (blogged on here) “found” 981,000 previously unreported jobs, about 750,000 of which were added between April 2005 and March 2006, with the rest coming between April and December of 2006. After BLS’s February 2008 revision, there should be some pretty decent evidence as to whether ADP’s methodology really is picking up on job activity every month that BLS is taking a year or two to detect. By February 2009, the jury should be completely in for the period ending December of 2007. I would expect that there will be a lot of ink consumed and bandwidth burned in the meantime attempting to compare BLS and ADP.
____________________________
UPDATE: ADP’s reported employment increase for February is 57,000. This report, plus the claims to better accuracy described below, build a bit of drama into Friday’s employment report from the government. Plus, there’s an additional nugget ADP will apparently start providing — a large employer/small employer breakdown:
Other new detail in The ADP National Employment Report shows that employment at small and medium size businesses employing less than 500 workers grew 86,000, while employment at larger businesses declined 29,000. Over the last six months, small and medium size businesses accounted for most of the growth in private nonfarm employment.
