Reclaiming Common Sense

February  ADP  Reported Growth in All Sectors.

The 2018 data was revised higher, so was January, "reducing" February Data

The ADP Payroll report is released the Wednesday prior to the monthly government Employment Situation Report. The headline ADP number is the seasonally adjusted payroll gains or losses. The headline Employment Situation number is the Non-Farm Payroll (NFP) number. The NFP data includes both private sector workers and government workers. This causes a difference to happen between the two headline numbers.

The article "Feb. 2019 ADP Report: Remarkable" covered a number of topics. The data underwent major revisions. The data was revised "all the way" to 2002. The upward revisions to the December 2018 and January 2019 data meant that over 200,000 jobs were shifted from February 2019 to the prior two months. The data revisions shifted the data between 2009 and 2010. The pace of growth for 2018 was changed. We saw growth in all sectors month to month and February to February. Most importantly, we saw our first 300,000 ADP Payroll Month since 2006.

There have been 483,000 payroll positions created during the first two months of the year. Some Commentators were lamenting the drop in ADP Payroll from 300,000 to 183,000 payroll positions. They forgot that the data for 2018 and January were revised higher by over 200,000 positions. If the upward revisions were stopped at 2018 then the 87,000 added to January would have been added to February and we would have been reporting a gain from 213,000 to 270,000.  Either way there would have been 483,000 payroll positions created during the first two months of the year.

President Trump recorded his first 300,000 ADP Payroll month with the revisions to the January data.  This means that he has more 300,000 ADP payroll months during his first two years and one month in office than President Obama achieved during eight years in office.The last time that we had a 300,000 ADP payroll month was during 2006.

Significant revisions were made to the 2002-2007 data.  This is important to note because it was during this period of time that former President George W Bush  had his Non-Farm payroll Streak. The sister site published an article "The Politics of Employment" when President Obama had a 47 month "job streak." President Bush had a 53 month jobs streak before the CES data. ta was revised to lower one month to a negative. It is worth noting that this revision dropped the November 2006 level by 53,000 positions thus giving that month a -13,000 payroll position number.

Significant revisions to the 2012-2017 data.  This was referenced in the prior ADP article. It appears that those years were not as good as originally reported. It was also referenced that the Great Recession was greater than originally recorded and the recovery during 2010 was apparently better than we experienced

Significant revisions to the Trade, Transportation and Utilities (TTU) data. There were also significant revisions to the Professional Business Services (PBS) data.  These revisions were noted with regard to the Current Employment Statistics (CES) last month with the release of the January Jobs Report.The "iceberg" chart looks significantly different than the one in the "Wages and Workers" article because that histogram examined the non-seasonally adjusted CES data and this one is the seasonally adjusted ADP data. It is important to note that both the PBS and the TTU data was revised lower for all of 2018.

The data was revised for all sectors between 2013 and 2018. The data for the PBS and TTU data was all the way back to April 2003.If you look at the data from January 2019 and February 2019 you will notice that every piece of data is revised back to 2003, and that all other sectors have the same value during March in the two data sets. This means seasonal factors were changed. It was only 2013 through 2018 where the March data, and the end of the year data, were completely revised.

What impact will this have on the comparisons between the ADP data and the CES data? This column has been critical on making comparisons between two different data sets. The ADP data is released two days prior to the Jobs Report data. The NSA CES  data is different from the SA ADP data, and it is understood that the seasonal factors for the ADP data and the CES data are probably different.  The January ADP data was revised higher, basically meeting the non-farm payroll data from the January Jobs Report.  People may ignore the difference between the headline NFP jobs data and the ADP Private Sector Data. It still seems as though the Jobs Report data should be over 230,000 private sector workers. The comparisons will still be made.

This was a good ADP report. We are off to a fast start. Next month there will be three data points from which a trend can start to materialize so that we can compare same month data from the various different years.  Almost all economic data sets have advance, preliminary, and final versions of the data. The ADP data  is similar. Almost all government data has annual revisions, sometimes going back years. The ADP data is similar. Watch the growth rates, the seasonal factors, and the revisions.

It's the economy.