Reclaiming Common Sense

The first jobs report of the month is the ADP Payroll report. The ADP report is released the Wednesday prior to the Friday Employment Situation Report, also known as the government;s jobs report. The ADP report measures "private Sector Payroll changes." The employment situation report measures private sector growth as well as non-farm payroll growth which includes government sector changes. The ADP data is all seasonally adjusted data. The employment situation report is created using seasonally adjusted (SA) from two data sets . Both of the government data sets have non-seasonally adjusted data available for analysis. Both the ADP data and the Current Employment Statistics (CES) data sets went through massive data revisions. The CES data set was revised between the December and January Jobs Reports. The ADP data was revised between the January and February Payroll Reports.

Will the ADP report indicate a growing or slowing payroll? There are a variety of ways to examine the data. How are the sectors growing from month to month? How are the sectors growing from March of last year to march of this year? How is the overall payroll level growing on a rolling year basis? The ADP data is a black box. We do not know what the underlying non-seasonally adjusted (NSA) data is. The revisions to the ADP data took the payroll changes from an upward trajectory to a flattening and possibly weakening growth curve.

Are we growing quickly or slowing that growth? We were seeing payroll expanding at a rate of over 2% prior to the ADP revisions. January 2018 growth was well over 2% while the previous January was well below 2%. The revisions to the data makes it look as though the best is behind us, during February of 2015. How can this be after a stellar first thirteen months in office for President Trump and for job creation? How can this be after coming off of a record February Jobs Report. Remember the ADP payroll data is a different from the Current Population Survey (CPS) jobs data which is different from the CES worker data. Different samples, different sample sizes, different seasonal factors. Different growth rates mean way different results. If we are growing closer to the 1.76% rate, up or down, the payroll number could be be a disappointing Payroll Number. Could it be under 100,000? No. March tends to be better than February. May tends to be better than March. July tends to be the best month of the year. What if we really were growing at over 2% a year? Could we see over 300,000 jobs added to the payroll? Over 400,000? Over 500,000?

The Month to Month data indicates that the lowest value that we should see reported is 189,000. We are not growing slower than March of 2005. last month the "rolling month" growth rate was 0.19%. This month to month growth rate is the same at March 2006. If we grow at this rate then we should see 233,000 new payroll jobs this month. February 2006 was the strongest growth rate recorded. The difference between a growth rate of 0.18% and 0.20% for the month to month growth is the difference between 226,000 and 251,000. The reason for the slight change between the 2006 and 2016 data in the accompanying table is the difference in the next decimal place.

The Month to Month data indicates that we should see growth in all but two sectors. The Information Technology sector  and the Trade, Transportation and Utility sector both could show declines. If they show improvement then this could be a huge report. Natural resources should see a large increase in payroll as the ground begins to thaw. The same thing is expected in construction. Financial services should see a spike in payroll due to the tax season. The leisure and hospitality sector should see payrolls climb in preparation for Spring Break, Memorial Day weekend, and Summer in general.

The March to March data indicates that all sectors except Natural Resources should be up over last March. All of the sectors except IT are already higher than they were last March, so this should be self-evident. Again, if IT Shows March to March improvement then this could be a huge report.

The Devil is in the data - the Seasonally adjusted data. How much will the prior data be revised after the major revisions last month? If the overall payroll number is revised higher for February by 25,000 that will reduce this month's growth by 25,000. If last month is revised lower, it will give a bigger boost to this month's number. This report does not have a non-seasonally adjusted data set that is available to the public, as do the CPS and CES data sets. The CPS and CES data are indicating huge gains possible during March. The CES seasonal factor has been sliding since 2009. The seasonal factor had been sliding between 1990 through 2002. A lower seasonal factor skews the SA CES data lower than it could be reported. The data for ADP only goes back to 2002, so we do not see the changes in the pre-201 data and the DotCom recession. We have had higher NSA CES annual, or rolling year, growth rates prior to other recessions. This data can proceed higher.

Are we growing at over 2% annually or under 1.8% annually according to the ADP data?  It depends. The CES data was revised for 2015 through 2017 with the release of January Employment Situation Report. The ADP data underwent similar revisions with the release of the February Payroll Report. "All" data in Governmentland have advance values, preliminary values, and final values. Unemployment dropped mid-February to Mid-March, according to the weekly continuing claims data, yet another data set. March should be a better hiring month than February, most years. Was February "too strong?" will the growth slow a little month to month because it was growing "too fast" last month? The participation rate has been improving as the economy has improved. The economy has improved as people have returned to work. What difference does the ADP number make? This question was asked by the predecessor to this column during 2014. A September 2014 article compared ADP to the State and Regional Data. It was found that the ADP number does not track well with the SA CES data nor the SA CES data during July 2014. This article is all about managing expectations.

Expect at least 215,000 people being added to the SA ADP payroll number this month. That number could be over 300,000 if the annual, rolling year, ADP growth rate is around 1.91%. We were growing over 2% before the revisions to the ADP data last month.  How high can these numbers grow?Will it be recorded? Will it be reported?

It's the economy.