So, today Automatic Data Processing Inc. (everybody calls them ADP) announced that they’ll be changing (their press release used the word “enhance”) the methodology they use on their widely followed monthly survey of private-sector hiring. For those of you who don’t know, ADP is a big payroll-processing company that also counts payroll numbers. ADP payrolls are important statistics to compliment the Bureau of Labor Statistics’ nonfarm payroll (NFP) data and have the potential to move the markets. Typically, the ADP report comes out the Wednesday before NFP info.
The problem – if you want to call it that – has been the criticisms that ADP’s data send out confusing signals. Lately, ADP’s estimates of private payrolls have been off when compared to private NFP. Last month, for example, ADP’s estimate of private sector job growth was 162,000. Compare that to the 114,000 new private sectors jobs estimated by the BLS (this estimate will of course be revised in the October and November reports). The new and improved ADP measure captured 88,000 new private sector jobs – closer to the BLS estimate.
What’s changing? ADP is now going to increase their sample size to better capture the population payroll estimates. With statistics, a larger sample size typically leads to a more accurate estimate of the true state of an overall population. In addition, the larger sample size means that ADP can capture more industries (moving from three to five) and companies of different sizes (also moving from three groups to five groups). That should mean a more robust and granular picture of hiring in the United States. Besides a larger sample, ADP is also tinkering with the formula to calculate the results. Oh, and they’re dumping Macroeconomic Advisors, their current data partner, for Moody’s.
I think it’s great that we’ll get a potentially more accurate picture of private payrolls from ADP. Improvements to statistics are always a good thing and give economists, policy makers, and all other data nerds a more accurate picture of the world. While 88,000 new jobs may not be as great as 114,000 new jobs, a more accurate estimate means that we’ll have better information that we can use to make better decisions.
If you’re interested, Marketwatch has a great graph that shows the differences between the new ADP, the old ADP, and the BLS estimates.