Jack Dunn - Reclaiming Common Sense

The Current Employment Statistics Data is the basis for the "Job Creation" numbers and the "Jobs Streak" that has been repeatedly discussed during the past three or four years, There are many ways to examine the data including Private Sector Growth or Non-farm payroll growth. Non-farm payroll does include government jobs. This column focuses on the Private Sector number because this is the focus of the President's "Job Streak." The seasonal factors used to convert the NSA CES data to the SA CES data change from month to month, seasonal to season, and year to year. The authors of th report have been using strong seasonal factors to artificially boost the SA CES Private Sector data. Last Month's Jobs Report Column detailed how:

  • The May Report, after the June Revisions, was a negative private sector number,
  • June was a head-scratcher.  We Added Non-seasonally adjusted (Current Population Survey)  jobs, losing 1 million part-time jobs and adding 1.3 million full-time job non-seasonally adjusted, and lost seasonally adjusted jobs. We also saw one of the strongest unadjusted worker growth months (Current Employment Statistics) for the month of June since 1998.
  •  July was Embellished.  The official number was skewed higher than it should have been by using an artificially high seasonal factor. We saw July Job Growth comparable to 2010and slower than 2013, 21014, or 2015.
  • We saw virtually no August CES Private Sector Growth during August. If the seasonal factor for converting the NSA data to the SA data was the same as August 2015 there could have been a contraction reported.
  • September we saw another skewing of the seasonal factor.If we used the seasonal factor from 2012 they would have reported 33,000 jobs added. Instead the number reported was
  • October we saw the seasonal factors skewed again. If we used the seasonal factors from 2015 used this year then only 25,000 Seasonally adjusted workers would have been added.

If the data is being skewed on a consistent basis do we really know where we are with regard to workers?

December and January are real worker loss months. Notice from the "No Worker Creation Crash, Yet" graph that every January we start in a worker hole. We spend the first four months of the year climbing out of the hole. We reach the job hiring penultimate peak during July or August, retreat, and surge through either November or December, We saw the largest growth in workers during 2014, then 2015, followed by 2013. We are growing at a slower pace than 2012 and better than 2010.

Growth has dropped below 2% for the past six months.  Worker Growth is needed for GDP growth and GDP growth is needed for Worker Growth.The last time we saw a June to November slide comparable to this was June 2006. Does this mean that a recession is looming? We will not know for probably six months. What we know is that if we are anticipating CES growth, it should be at a NSA rate lower than last December.

We have seen growth during the past two Decembers of 0.10% and 0.08% the past two years. Normally, 2000-2011,  we lose workers. Have we been gaining workers because the CES data double counts people working two jobs? Have be seen NSA CES growth because we have the same amount of work being accomplished by more part-time workers instead of fewer full-time workers?

The Bouncing Ball Seasonal Factor. Sometimes we see a straight-line trend on the seasonal factors and sometime we see a scatter plot.  The trend for the past four years has been downward.If the seasonal factor used to convert the NSA CES private sector worker number is greater than 0.9955 we will know that they are skewing the data.

Will we have Real Worker Contraction and Fake Worker Growth? The combinations and permutations are not quite endless this month. We should see some level of NSA Private Sector Worker Contraction reported next week. The "Contraction table" allows for some growth. We have seen contraction thirteen of the past sixteen years. We are not in a 2008 or 2009 environment. We are not in a 2014 or 2015 environment. Expect modest NSA CES contraction and an optimistic seasonal factor.

Expect a SA Private Sector number under 171,000. Do not be surprised if it is closer to 100,000. The economy is slowing.

Its the economy.

The most anticipated economic number during the course of any given month is the "Jobs Report" number.  The Jobs Report, or Employment Situation Report, is created from two discreet data sets: The Current Population Survey (CPS) data and the Current Employment Statistics (CES) data. The CPS data measure jobs, full-time and part-time, plus the number of unemployed, plus the non-institutional workforce population. The unemployment rate and participation rate are generated from the CPS data. The CES data measures workers. Each data set has their own sample size and seasonal factors. The CPS data has unique  seasonal factors for full-time jobs, part-time jobs, and the unemployed. The seasonal factors change by month and year. When seasonal factors are higher or lower than the historical range FACTs (False Assertions Considered to be True) are created.

The Mother of all FACTs is the Consecutive Months of Job Creation Streak. The Job Streak is the seasonally Adjusted Job Streak, using the Private Sector CES worker data. The administration is touting 15.6 Million jobs added over 81 months - note that the no longer state consecutive months. We lost private sector workers during June - just 1,000 - still the streak ended.This author has analyzed the data for years. The streak did not begin when they said it did.The streak began during March 2010 and ended during January 2015. A downward revision to the January 2015 data, actually all of the 2015 data, spurred a revision backwards "for as far as the eyes could see." President Bush had a 53 month streak of Non-farm payroll growth that was never reported. This was not the first time that this column addressed the Streak.  "President Obama's Streak - When did it Begin, When did it End" showed that the streak did not begin until February of 2011.  If the streak began during February 2011 and ended January 2015 it was a streak of 60 months. It is also important to note that worker growth has been slowing and may be heading into a worker recession. A "worker fade" may be under way.

What should we expect this month? Job Contraction should be recorded, job contraction could be reported, again. Worker contraction should be recorded while worker expansion may be reported. It is almost a certainty that the "worker" number reported will be positive until the day that Obama leaves office. What is recorded is the NSA data. What is reported is the SA data. This column will dig into both data sets and project growth or contraction by category and seasonal factors by category.

We should see real job loss recorded and possible seasonally adjusted job contraction reported. The media has started to understand the existence of the two data sets. They have started to wrap their brains around the possibility that job loss, seasonally adjusted, could be reported and private sector worker growth could be reported. We tend to lose full-time jobs and gain part-time jobs during December. There is a huge surge in workers leading up to the middle of December when the surveys are taken. Companies may be performing seasonal losses or shutting down for good prior to the end of the year. We have seen both NSA full-time (FT) job and part-time (PT) job loss during December 2010, 2011, 2012, 2013, and 2014.Last year we saw unadjusted FT job gains washed away with part-time job losses. The only December since December 2003 where we saw net real, unadjusted, job gains was December of 2006 - Right before the beginning of the Housing Recession.

The Seasonally Adjusted Jobs Changes Should Report FT Job gains and PT Job losses.  We experienced real and fake FT job losses during December of 2007, 2008, and 2009.It is very possible that the media on television will be scratching their heads trying to figure out how we added over half a million full-time jobs and only added 150,000 workers. They will explain it away with a discussion on the surge of dual job workers. (This will be covered in another column.) That is an easier solution than the real answer that he two data sets are only related by the date that they are released. Depending upon the seasonal factors used, we could record half a million combined NSA  jobs lost and report half a million combined SA  jobs added.

We normally see the seasonally adjusted an non-seasonally adjusted unemployment numbers march in lock-step.If the NSA U3 number falls the SA number falls. If the NSA U3 number increases then SA number increases. Last December we barely saw any change in the unemployment level between November and December.

Unemployment levels tend to drop during December only to spike during January. We have seen unemployment claims, first-time and continuing claims, creep up, non-seasonally adjusted, during the past few weeks. The continuing claims number has edged up by 250,000 since mid-November. Just as the CES and CPS data sets are separate and unique, so are the CPS data and the weekly claims data. We may or may not see a jump of 250,000 unemployed workers because the covered insured is so low as a portion of the unemployment rate. Only 1.5% of the population that are unemployed (under 5% U3) are covered by unemployment benefits. Could we see the unemployment claims increase by 80,000? The official unemployment number could go up slightly, drop slightly, or sort of slide sideways as it did last year. What is obvious is that this is the calm before the storm. Unemployment will spike during January. It's a tradition.

Everything should drop. The Seasonal Factors are all over the place. The best case scenario might be to have full-time jobs, part-time jobs, and the unemployment levels remain unchanged. It appears that we will see full-time jobs drop, unemployment drop, and the population grow. A growing population and decreasing participants lead to a falling participation rate and a falling unemployment rate.

The Most important thing to take away from the CPS data is that down is normal and that up may be reported. we need   203,000 jobs created to keep up with the 0.08% population growth from 254,540,000 to 254,743,000 in workforce population.

The next section will examine the CES data. Just as the seasonal factors for the CPS data are all over the place, it will become apparent that the CES seasonal factors are all over the place, too.