The February Jobs Report - A Tale of Two Data Sets
The ADP report was released earlier this week to faint praise. The ADP Payroll report is often seen as an indicator as to what to expect from monthly government jobs report. The data is from different data sets with different sample sizes, different seasonal factors, and different margins of error. Fruits of a different color and type. Others make the comparisons while I explain how that is not the best thing to do. It is the battle of the data sets.
The ADP forecast was for growth in all sectors February to February and all except sectors month to month, seasonally adjusted. There were questions on how strong the growth would be and how much the data would be revised. The article "Feb. ADP Should Be Strong, Growth in All Sectors" expected a number around 230,000 and possibly as high as 269,000. The question was how much would the prior data be revised.
The ADP report had revisions to all of the available data back to 2003. This is significant because the data for 2018 was revised higher than it was reported just last month. The article "Feb. 2019 ADP Report: Remarkable" detailed how the 2018 data was revised 122,000 payroll positions higher than last reported. It also explained how the January data was 209,000 positions higher than reported last month, or a next upward revision of 87,000 payroll positions, as was generally reported in the news. The upward revision to the January data gave us our first 300,000 jobs number since the Great Recession. It also turned a potentially 392,000 payroll number to a 183,000 number. On the upside, we did see growth in seasonally adjusted payroll positions month to month and February to February in all sectors.
The February Jobs Report Forecast Article projected strong seasonally adjusted worker gains, too. The article "February 2019 Jobs Report Forecast: Big League" examined month to month and February to February changes in the Current Employment Statistics(CES) worker data and the Current Population Survey (CPS) jobs and unemployment data. It started with a look at the annual growth projection and found that February tends to grow faster than January. Based on this data it was thought that we would exceed the 2.15% growth rate recorded during January and possibly hit 2.20%. This would me that we could add between 275,000 and 338,000 SA Private Sector Workers, before revisions to the January and December data were made. All sectors were expected to add workers except possibly Information Technology (IT.) We have seen strength in IT during recent months so there was the possibility for a slight growth in February to February IT workers.
The month to month forecast was for grow month to month in all sectors except IT and "Other Services." Once again, there was some thinking that IT could grow month to month based on recent changes in the sector. The seasonal factors have been declining since 2010. If the seasonal factors dropped as low as expected, and if the growth rate was not as strong as last year's record setting growth rate for the month of February, the thought was that we could see month to month growth of 200,000 to 235,000 before revisions to the prior months data. This lower month to month trajectory effectively placed a "lid" on month to month growth at 235,000.
That said, the Current Population Survey Data was projecting a massive spike in full-time and part-time jobs, and a drop in unemployment. This is one of those months where the worker data is significantly different from the CPS jobs data. The non-seasonally adjusted (NSA) CES private sector data was looking to add 700,000 to 800,000 workers whereas the NSA CPS data was projecting the possibility of 1.0 million to 1.5 million total jobs being created. Some of this spike was expected to be offset by a drop in unemployment. The seasonal factors could turn massive FT and PT gains into potential losses for full-time jobs and gains for part-time jobs. Participation and unemployment data were expected to report improving conditions.
What the heck happened?
We added more Non-seasonally Adjusted Private Sector Workers than 2014, 2015, and 2016 and reported adding fewer workers. The Headline Non-Farm Payroll data was reported at a seriously small gain of 20,000. Private sector workers were reported at 25,000 workers being added during February. This is much lower than February 2014 when we added 157,000 workers, and lower than February 2015 when we added 227,000 private sector workers and 2016 when we added 206,000 seasonally adjusted private sector workers. This February we added .418,000 NSA CES workers this month, compared to 310,000 during February 2014, 396,000 during February 2015, and 398,000 during February 2016. We recorded more workers added and reported fewer workers added this February.
We saw month to month growth reported as 0.33%, after the revisions to the December and January data were made. The non-seasonally adjusted December data was revised up by 41,000 and the January data was revised up 94,000, or a net 53,000 from the December revisions. This means that the February data should have been reported at least 41,000 jobs higher, and possibly 94,000 workers higher. The growth rate is comparable to 2015 (0.34%) and 2016 (0.33%) and significantly better than 2014 (0.27%.) If we use the seasonal factor from February 2015 then the headline number should have been 194,000.
How fast is the economy growing? Are workers growing close to 2.00%? This February the private sector workforce grew at a faster annual pace than February 2017, faster than February 2018, and yet slower than January this year. Remember that this is the private sector data.
Up is Down in Governmentland, using Government Math. The private sector data recorded a jump of 418,000 workers while the non-farm payroll jumped 827,000, meaning that we had 409,000 NSA government workers join (rejoin) the workforce. Remember the Government Shutdown? The 409,000 government workers reduced the NFP number by 5,000 workers. Up is down.
The CPS Jobs Numbers paints an entirely different picture. The CPS data is used to calculate the participation rate and the unemployment rate. We saw 670,000 NSA Full-time (FT) jobs added and 532,000 NSA Part-time (PT) jobs added during February. This means that 1.202 million jobs were created last month. This is more than were created during 2014, 2015, 2016, and 2017. February 2018 was the all-time high for the month of February.
If you compare the 2014-2016 data and the 2019 data Up is Down again. The non-seasonally adjusted data for CPS jobs were:
Now compare the Seasonally Adjusted data:
We had 180,000 more combined NSA jobs this February than February 2016 and reported 197,000 FEWER combined jobs.
How low can the unemployment level fall? This February we had 6.625 million non-seasonally adjusted (NSA) U-3 unemployed workers. We had 6.563 million during February 1999, 6.284 million during February 2000, and 6.523 million during February 2001. The thing to remember here is that we had a workforce population of 206.873 million during 1999, 211.576 million during 2000, and 214.110 million during February 2001. This year we have 258.392 million workers. We have 100,000 to 400,000 more unemployed workers than we had during 1999-2001 for 44 million to 52 million more workers.
We are missing 10 million workers. If we compare the workforce participation rate that we had during 1999 with what we have this February then we are missing 9.827 million workers.
There is a large amount of information in this report to unpack. What happened with wages and workers? How is President Trump doing compared to former presidents Reagan, Clinton, George W. Bush, and Obama? How are men and women doing? What is happening with multiple job workers? Is the workforce still graying? The CPS data factors into some of these articles while the CES data factors into other articles. It's the tale of two data sets. This will all be covered next week.
It's the Economy.