Up is Up - Jobs Surge
The monthly "Jobs Report, officially known as the Employment Situation Report, is created using two different databases: The Current Employment Situation (CES) data, or the Establishment Data, and the Current Population Survey (CPS) data, or the Household data. The CPS data measures full-time jobs, part-time jobs, unemployed workers and the workforce population. It is from this data that the unemployment rate and the participation rate are calculated. The CES data measures worker levels and is used to report the level of workers in the various Super Sectors. There are a number of ways to examine the data. This column write an article in preparation for today's release titled "September Jobs Forecast: Down is Up."
We tend to lose non-seasonally adjusted jobs and workers every September as the Summer ends and some of the seasonal workers return to college. We also normally see the non-seasonally adjusted level of unemployed workers fall as those people are not actively searching for jobs, therefore they are not unemployed.
There has also been a considerable amount of time, energy, and effort projecting the impact of Hurricane Harvey, Irma, and Maria on the jobs and unemployment numbers. Hurricane Harvey hit after the August Jobs report data collection date and before the September collection date so Harvey's impact will be measurable this month. Note that this column has written articles on this subject and that Tropical Storm Allison had a modest impact from August of to August 2002. There has been research done on the Texas, Florida, and Puerto Rican Job Markets prior to the release of this jobs report.
Contrary to what you have heard, we recorded job gains last month. The CPS Jobs data recorded and "reported" job gains this month. We saw over one million non-seasonally adjusted part-time jobs added this month. This was not quite as many part-time jobs as were added last September, and it was more than September 2013-2015.
The could have reported the best September Jobs data since 2003. Normally the SA CPS jobs data "reports" seasonally adjusted job losses. We lost jobs September 2003-2005, 2008-2010, and during September 2015. This was a strong "Jobs" Report.
We recorded a large drop in Unemployed workers and reported a large drop in unemployed workers. Some years we have seen a drop in real, non-seasonally adjusted, workers only to report an increase in unemployed workers. This year we saw the third largest drop in unemployed workers for the month of September since 2003.
We saw participation record a higher rate even as the unemployment rate fell. Unemployed workers are participants. An unemployed worker counts the same as a full-time job or a part-time job. We saw the participation rate record an increase from 63.00% to 63.02%. The SA participation rate rose from 62.84% to 63.05% The September NSA participation rate is higher than it was during September of 2014, September 2015, and September 2016.
We are not at full-employment. The common definition of full-employment is an unemployment rate under 5.0%. The NSA U-3 unemployment rate was 4.07% this September compared to 4.37% during September of 2006. The problem is that the participation rate right now is 3.07% lower than it was during September of 2006. This means that there are millions of "non-participants." They are neither employed nor unemployed. We are missing 7.864 million jobs or unemployed workers. If these missing workers/jobs are added to the unemployed ranks the "real" unemployment rate is 8.54%, using the September 2006 unemployment rate as the "zero" missing participant level. The September U-7 unemployment rate peaked during 2011 at 11.45% when the official unemployment rate was 8.78%. The missing participant level has been on a fairly steady decline since 2011.
The net takeaway for the CPS data is that more people are working, less people are unemployed, and the participation rate continues to improve. What about the official number that showed net worker loss?
Down is Down, Workers Decline (End of Summer.)
This is the tale of two data sets. The first have of this column addressed the CPS Household Data. This section is going to address the CES Employer Data. We saw a contraction in the non-seasonally adjusted CES worker data. We saw the largest month to month drop, by percentage, during the month of September that we have seen since the great recession. It was almost as bad as the drop in workers recorded during September 2001.
Normally the seasonal factors convert Monthly Worker Losses into Monthly Worker Gains. We saw worker contractions during September 1981, 1982, 1990, 2001, 2002, 2008, 2009. Three of those years were the first Septembers in office for new Presidents (1981, 2001, 2009.)
It could have been reported worse than it was reported. The seasonal factors have been creeping higher every September for a very long time, as was detailed in the September Jobs Report Forecast article. If last year's seasonal factor was used the number would have been reported even lower with a loss of 43,000 workers. If we used the seasonal factors from 2013 then it would have been reported at a loss of 137,000 seasonally adjusted private sector workers.
This data requires further examination. This column "always" digs into the jobs data after the rest of the world appears to move on to whatever the "crisis du jour" is for that day. This column will address the multiple job worker data (which rose,) the comparison of five Presidents after Eight Months (Reagan, Clinton, George W. Bush, and Obama,) the Super Sector data (which appears to show Leisure and Hospitality declines negating and reducing real worker gains,) as well as the changes in employment for men and women and the changes in employment and unemployment for all people based on age group.
This jobs report was a tale of two data sets. One data set recorded and reported real job growth. One data set recorded and reported worker declines. It is possible that the unemployment rate fell because the Leisure and Hospitality Sector have an elevated level of people who are working part-time. The multiple job worker data may show a spike in the number of people working two part-time jobs or those who work a full-time job and a part-time job. The problem is that when the data gets difficult many people ignore the data and rely on the report. The State and Local Data Employment and Unemployment Data will be released October 20th, at which time the data for Texas, Florida, and Puerto Rico can be examined in more detail.
It's the economy.