Are ADP Jobs Numbers Perfectly In Sync With NFP Private Payrolls Data?

Those bracing for upbeat Labor Department non-farm payrolls, or NFP, data for March following an exceedingly strong Automatic Data Processing ADP private payrolls report were in for a disappointment.

The March NFP Report: A Dampener

The Labor Department's establishment survey, which is used to calculate payroll numbers, showed that the economy added merely 98,000 jobs in March. Additionally, there were downward revisions to January and February numbers.

For the whole of 2016, the economy added 2.24 million jobs, with the average job gains for the year at 187,000 jobs.

Retail jobs contracted by 30,000, while the construction, manufacturing, leisure and hospitality and healthcare and social assistance sectors saw a notable slowdown in job growth.

The NFP report's customary release date is the first Friday of every month, and the March report was released on March 7.

Rewind to that Wednesday, March 5. The ADP report, a private initiative, showed an addition of 263,000 to private payrolls in March. Private payrolls of the Labor Department survey account for roughly 90 percent of the total payroll gains, excluding government jobs alone.

The ADP report is released the first Wednesday of each month and comes two days ahead of the NFP report. The report is considered a precursor to the NFP data.

ADP's March non-farm private payrolls number is in stark contrast to the 89,000 private payroll gain forecast for March by the Labor Department's survey.

Survey Methodology: NFP Vs. ADP

The Labor Department survey done under the Current Employment Statistics program surveys about 147,000 businesses and government agencies, representing 634,000 individual worksites. Data gathered includes employment, hours and earnings.

Meanwhile, Moody's Analytics, which partners with ADP in releasing the ADP private payrolls report, gets the ADP payroll files, given that ADP client base includes about one-fifth of the U.S. private payroll employment. The data from the ADP payroll files are compiled into time series, beginning in 2001. The data is then processed over a number of steps, which includes:

• Removal of extreme values.
• Identifying clients by industry and company size in order to aggregate the individual client data into industries and company size.
• Creation of matched pairs.
• Seasonal adjustment.
• Adjustments to match the industry and size distribution of the BLS's employment data.

Historical Correlation Between Both Surveys

An analysis of how these two data are correlated going back to three years in the past did not inspire much confidence despite Moody's analytics claims that it adjusts the ADP data to match the CES distribution.

Source: Y Charts

Hard data analysis shows that of the 36 months of the three years we took into account, only eight months showed the numbers falling within 10 percent, plus or minus range. The change in either direction was between 10 and 20 percent in about eight instances. The variation of 20–50 percent was the most prevalent, with 16 instances. In two cases, the change was between 50 and 100 percent and in another two instances, over 100 percent.

The Labor Department data is subject to revisions sometimes even three to four times.

What Can Cause The Difference?

The Labor Department survey takes into account all employees who are actually paid, while the ADP survey counts all employees listed as active on an employer's payroll. Therefore, temporary labor force changes such as strikes or structural changes can impact the numbers.

Therefore, ADP may not be an accurate predictor for the month's NFP numbers. Traders cannot position themselves for a jobs report Friday, just with the knowledge of the ADP numbers. Roughly one-third of the time, there has been discordance even with the direction of the numbers. That said, in isolation, economic data points are what matters.

Related Links:

The Markets Reel On Weak Jobs Data And U.S. Strike On Syria

17 Must-Watch Metrics In The New Trump Era