No surprise, then, that the company recently revamped its formula for assessing the labor market and partnered with Moody's Analytics in hopes of improving the report's accuracy.
In today's release, the first one using ADP's new methodology, the payroll company said non-farm private employers added 158,000 jobs in October. That's in line with the 162,000 jobs it said the economy generated in September.
Or not. According to ADP's new formula, job growth in September was half of what the company previously reported:
Private sector employment increased by 88,200 jobs from August to September, according to the September ADP National Employment Report... Last month's employment estimate was revised down from 80,000 to 76,400 jobs.
In gauging the state of the job market in a given month, the difference between 162,000 and 88,200 jobs is a major difference -- and that's not the only downward revision in the report. ADP previously said that the U.S. added 4,000 manufacturing jobs in October. But the latest report now says the economy actually lost 17,400 manufacturing jobs.
Despite such discrepancies, ADP expresses confidence that its monthly reports "will further align" the company's findings with the U.S. Labor Department's monthly job figures. Historically, the ADP and Bureau of Labor Statistics jobs data have often been significantly different, causing confusion over the pace of job growth and occasionally catching investors by surprise.
It's odd that the entire point of this revision is to make sure ADP comes up with the same numbers as the government does. That's very different than trying to give better, more accurate information.
If ADP's new numbers are correct, then it stands to reason that the old ones were wrong. ADP won't admit this, at least not directly, because it would put the company's broader competence into question. The closest it comes to acknowledging the problem is this convoluted answer in an FAQ on the company's website.
What if any impact does the increased size of the data set have on year-over-year or other comparisons with the historical data?
The larger data set, with the new methodology by Moody's Analytics, will help enable ADP Research Institute to more closely match the final BLS numbers.
Financial blog Zerohedge applied ADP's new methodology to all of the company's job reports to date for 2012 and found it had overstated the number of new jobs by 419,000. While that is clearly bad news for the economy, it is offset a bit by the fact that, according to analyst Mish Shedlock, ADP under-reported the number of jobs created in 2011 by 229,000.
ADP's partnership with Moody's (MCO) in developing its revised methodology is also worth noting. You'll recall that Moody's is one of the three big credit rating agencies. All three of those agencies have admitted that, in the years leading up to the housing bust, they erroneously gave AAA ratings to mortgage-backed securities, usually at the behest of Wall Street banks that were selling the securities. All three have also consistently been months behind the curve in changing their ratings on sovereign debt issues.
In other words, it is anything but certain that ADP's teaming with Moody's will improve the accuracy of its job reports.