Jack Dunn - Reclaiming Common Sense

What Will Be The Headline Number for September? The Headline Number is the Seasonally Adjusted (SA) Current Employment Statistics (CES) number. More importantly,  it is the SA CES Private Sector Number. Sometimes the media is discussing the Private Sector number in one breathe and discussing the non-farm payroll number in the next sentence. Non-Farm payroll includes the government sector - Total Private sector, obviously, does not include government workers. The CES correction table reveals that non-farm payroll, non-seasonally adjusted, should increase over last month. How can we lose jobs and gain workers? The data is from two different samples with two different sample sizes.The difference is between Private Sector workers and Non-Farm workers.


We have lost NSA Private Sector workers every year since 1998. How do we have a "streak" when we lose workers every September? They seasonally adjust out the reality of annual cycles in employment. The good news is that we have seen smaller NSA CES private sector contractions during September 2014 and 2015. than September 2006 and 2007. Are they trending downwards?  Will we see more worker loss this month than we saw the past two Septembers?


We Normally See Seasonally Adjusted CES Worker Losses during August and September during Recessions. We saw SA CES private sector losses during September of 2001 and September 2002. We saw SA CES private sector job losses during September 2008 and September 2009.


We should have had SA CES Private Sector Job Worker Loss during August. We started this column with the mention of last month's CES data and how it should have been reported as a negative value. If they had reported a contraction it would have been the first time a contraction was reported in the mainstream media. This column attempted to break the story that we saw worker loss during May when the June jobs report was released. They revised the May data down for the private sector and the non-farm data. The non-farm number was still positive, so nobody thought much about the situation.


If the data from prior months is downwardly revised there may be an artificial bump in the jobs created number for September. The data that will be released this Friday will be the advance data for September, the preliminary data for August, and the final data for July.  If they drop the basis for the starting point by 50,000 workers then a loss of 10,000workers could be reported as a bump of 40,000 jobs.


Seasonal Factors Matter. Last month the seasonal factor used to convert the NSA data to the SA data were manufactured to keep the data reported as a positive number. The seasonal factor used was never previously used. The seasonal factor used was not considered by this column as a viable seasonal factor.


If we Lose jobs as we did during 2001 or 2008  or 2009 we should see a SA Worker Loss Reported.  Even if we see the worker contraction we saw during 2013 or 2015 we should see a "Jobs number"  that is reported under 150,000. The problem is that the worker contraction rate can be masked by the seasonal factor.


The Seasonal Factor has been creeping. Pick your growth rate, or in this case contraction rate, and pick your seasonal factor, and you can get entirely different results. If we lose jobs like we did during September 2002 and use the seasonal factor for 2012 we will lose 7,000 SA workers. If we use the seasonal factor from 2015 that same 2002 number would be published as a gain of 80,000. If we contract at the same pace as we did during 2013 and use the seasonal factor from 2013 we should gain 143,000 workers. If they use the next seasonal factor in the progression, the ??? rate,   we could contract like we did during 2009 and almost add jobs.


Is this a plateau or a plummet in the making?  We saw a cresting in jobs for the year during August of the years 2001, 2002, and 2008. Will we drop down to the annual levels of job creation seen during 2007 or 2013? This report could be critical for the direction of the economy politically and economically. A downward turn that is recorded or reported could impact the elections. There are so many early voting states that this could be the last report to swing voters.


If I were interviewed this week what points would I make?

  • Focus on the Private Sector Number. The President's Job Streak was based on the private sector number. Ignore the Non-Farm number. Governments could go on a hiring spree to boost the non-farm numbers.
  • Dig into the Jobs Created and Lost since July of 2007. The President and his surrogates love to discuss the number of "jobs" created since the bottom of the recession. They fail to recognize the jobs lost between July 2007 and January 2010.We had to lose 15 million jobs to add 20 million jobs.
  • Dig into the changes in full-time and part-time jobs numbers. We added part-time jobs during the recession as we shed million of full-time jobs. We did not start adding full-time jobs until last Summer - 8 years after the beginning of the jobs recession. We lost those full-time jobs last Fall and Winter. We only just started adding full-time jobs to the 2007 level this Summer - Nine years after the start of the Jobs Recession.
  • Comment on the changes in the Jobs Sectors - non-seasonally adjusted, of course. Last month we saw five sectors that had not recovered from the Jobs Recession. Will we see contraction in the same industries this month?
  • Comment on the Workforce Participation Rate - and Explain it clearly. We have seen a slide in the workforce participation rate during the past 8 years. We have a failure to participate. Participation was down for virtually every age group last month, other than those over the age of 64. The effectyive unemployment rate when participation is factored into the equation is over 10%.
  • Disregard whatever the ADP report reveals. This column has analyzed the differences between the ADP report and the Employment Situation report. All of the ADP data is seasonally adjusted. How do they know what the government's seasonal factors will be?


Why should you trust this column? I am watching "Varney and Company" as I draft this column. Recently Stuart Varney poo-pooed forecast columns. I was a saleman for 15 years. I knew what was possible and what was impossible for me to sell to a customer. I managed his/her/their expectations with regard to what kind of home I could find in their respective price ranges or communities. I told them to expect hiccups in either the home search, bidding/inspection, or closing of the property. If there were no hiccups they might appreciate it or think I was overpaid. One or two hiccups and they would think that it was okay and that I earned my commission. If there were three hiccups and we managed to close the deal they might think I was overpaid or that I had really worked for my paycheck. I would ask them where they wanted their hiccup. They would say "We don't want any hiccups." I would say "So if we don't have any hiccups you won't complained that I was overpaid." Some would say to me "We won't go that far" with a grin or two.


This column and all the articles written herein are there to manage expectations, to explain what really happened, and hopefully to calms nervous minds. We hit a hiccup during May. We hit a hiccup during August. We might have some hiccups and indigestion this month.


Hang on. Fasten your seat belt. Place your trays in the upright and locked position. We will be experiencing some turbulence for the next two Jobs Reports.


If you want to catch up on the Jobs Report Columns for this year:







 

September "Jobs" Report - Seasonal Factors Matter


The most anticipated number of the month is the Official Seasonally Adjusted Current Employment Statistics Worker Number. This number is called the "jobs number." The Employment Situation Report is often referred to as the "Jobs Report." It needs to be remembered that there is data that is released with the report, that the data is seasonally adjusted month to month, year by year, category by category. The headline "Jobs Number" is actually the "Workers Number." There is a huge difference between the two.


This column has published articles that have demonstrated that the President's Job Streak did not begin when they said it did and that it ended this past May. If the Bureau of Labor Statistics had stopped at revising the prior jobs data at 2015 then the streak would have ended January of 2015. They had to revise the data back to 2005 to maintain the streak. The streak would have begun February of 2011 and ended January 2015. Four years - forty eight months. They should have reported a CES worker loss last month. e lost over 1 million part-time jobs last month. Nobody else commented on it because that was the non-seasonally adjusted jobs (Current Population Survey)(CPS) data.


This column will projected the possibilities for the CPS Data and the CES data. The CPS data measures the number of full-time jobs created/lost. the number of part-time jobs created/lost, the changes in the number of unemployed workers, and the changes in the workforce population. This data is used to generate the unemployment rate and the participation rate. The CES data is used to create the President's "jobs" number.


We "Always" Record Full-time Job Losses and Part-time Job Gains during September. The first histogram reveals that we lost net September Jobs nine of the past thirteen years. Yes, We added more part-time jobs than we lost full-time jobs during 2012-2014. Last year we lost a quarter million jobs during the month of September.   This data is the recorded, non-seasonally adjusted (NSA,) data.


We Normally "Report" net job losses when the CPS data is actually reported.  The seasonal factors used to convert the Full-time jobs data can swing a loss into a gain. Likewise, the seasonal factors used to convert the Part-time jobs data to the seasonally adjusted data can turn gains into losses.  Up is down, down is up.


Unemployment tends to be Recorded and Reported as declining during August. Some people  find jobs. Some people run out of benefits - those people who were unemployed during January normally only receive 6 months of benefits. These people can extend those benefits briefly by finding temporary work and then can restart the remainder of their benefit period. Some people go back to school.  Here is what is important to remember: Unemployed workers are counted in the participation number. Participants = FT + PT + U3. The participation rate is equal to Participants divided by the workforce population.


Will the Full-time Jobs Lost this Month be Covered by Unemployment Insurance? If you look at the first histogram, we "always" lose full-time jobs and gain part-time jobs, when examining the non-seasonally adjusted data, also known as reality. It will be interesting to see what happens because we have seen a drop in the continuing claims number and the first-time unemployment claims numbers in the weekly unemployment claims numbers. You will notice in the first graph that we saw more part-time jobs added than full-time jobs were lost during 2011-2014. You will also notice that we had net job losses during August of last year. The graph below the non-seasonally adjusted job histogram reveals that we lost seasonally adjusted jobs last August. The streak does not exist because the White House has been cherry picking the CES worker numbers and has not been using the CPS jobs number. More on that later in this column.


The non-seasonally adjusted will be manipulated - Up is Down, Down is Up. Still will end up with a Drop in Participation. This column has published many articles on the drops in the participation rate and how it is artificially deflating the unemployment rate. The "Four Presidents" columns go into detail on how the effective unemployment rate is certainly being under-reported.  The tables below  this paragraph show that We should record full-time job losses and report full-time job gains. We should record part-time job gains and report part-time job losses. The number of unemployed may decline both in the non-seasonally and seasonally adjusted formats. When these job losses and drops in unemployment are combined the participation rate will most certainly fall. This is not good. It is worse that this is not being explained elsewhere. We have seen the highest level of participation that we will see this year.


How will the data be seasonally adjusted/manipulated? Will Will Grow the Population. We Will Lose Full-time Jobs. We Will Gain Part-time Jobs. We will see Unemployment rise. Rarely is the data ever this clear. The seasonal factors used to covert the non-seasonally adjusted data to the seasonally adjusted data swings from category to category and year to year.


Could we record a loss of 1 Million participants and report an increase in the participation rate? The recorded rate is the non-seasonally adjusted data. the reported data is the seasonally adjusted data. We could lose over 1 million full-time jobs and approach two million . We could lose half a million to a million unemployed workers. The tables show you how the data is manipulated.