The Devil is in the Seasonal Factors
Last month we received a solid ADP report and middling Employment Situation report. The ADP data is seasonally adjusted and reported 250,000 jobs added. The Current Employment Statistics (CES) data is recorded as a non-seasonally adjusted (NSA) and reported as seasonally adjusted (SA) data. The seasonally adjusted data was reported at 146,000 when it could have been reported at 256,000 workers. The seasonal factors that were used to convert the NSA CES Private Sector data were lower than anticipated which lowered the reported CES data. The ADP forecast for January is for another strong report. The strength is based on normal month to month growth in seasonally adjusted jobs and in the total annual growth rate. A similar approach is used to project the potential growth in the CES data. The bad news is that some of you lost work last month. The good news is that they were normal seasonal losses. The good news is that this data will be seasonally adjusted a gains and that the gains may convince employers to hire more people.
The Headline Number is the Seasonally Adjusted CES Worker Number - and That may spike much higher. last January we had one of the smallest drops in the NSA CES Private Sector worker level that we had recorded during the prior decade. The normal drop is around 2.00%. Last year it only dropped 1.96%. Thankfully the headline number is a seasonally adjusted number. Whereas "every January" we lose NSA CES workers, almost every January we ad SA CES workers. Last year's NSA contraction of 1.96% was reported as a, n expansion of 0.17%. The worst January SA data was reported during January of 2008, 2009, and 2010. We saw virtually no change in the SA CES data during January 2008 and January 2010.
The seasonal factors are all over the place - that is an economic technical term. If you look at the seasonal factors used to convert the NSA CES data to the SA CES data it appears that someone threw a data at the wall and took one step over each year. Some months there is a steady pattern up or down. January is not one of those months. The seasonal factor last year was high - trying to boost the final jobs number for President Obama. This is going to be President Trump's final jobs report for his first year because the collection date is January 12th and the inauguration was January 20th.
Growth, seasonal factors, and adjustments make a difference. We already know that the growth is actually a contraction that is seasonally adjusted. If we see a similar rate of change as last year, and a similar seasonal factor, then we could add over 300,000 private sector workers. I am not anticipating as favorable a seasonal factor. If we grow as we did during 2014 or 2015 and use the 2014 seasonal factor then add 232,000 to 264,000 private sector workers. If we see 25,000 workers added to last month's official value that would "borrow 25,000" workers from this month's data. Instead of a solid 232,000 to 264,000 that reported value could be 207,000 to 239,000.
The Annual Growth rate could be over 1.8%. The first number calculated was done using December to January growth.If we look at the annual growth rate, where we have been, December 2017 was 1.73% compared to January 2017 at 1.68%. If we "just" grow at the same rate as this past December we will see 204,000 workers added. If we grow at December's annual rate of 1.73% we will see 266,000 seasonally adjusted workers added. If we grow at the rate we did during January of 1996, 1.81%, then we should see 364,000 workers added If we grow at the rate we were during January of 2014, 1.87% we should see 446,000 workers added. We should see those numbers, based in the 2014 seasonal factor. The question is how low will the seasonal factor go? How high will the revisions to the December data be? We will not hit the level seen during 2012, 2014, or 2015. That would be a gear stripping jump in annual growth. Expect a number over 250,000 and most likely around 265,000 seasonally adjusted private sector workers added.
The rest of the story (data) comes from the Current Population Survey Data
The CES data is the headline data for "jobs" even though it measures workers.The true jobs number is derived from the Current Population Survey data, or household data. This data measures the quantity of full-time jobs, part-time jobs, unemployed workers, and workforce population. This data is then used to calculate the unemployment rate and the participation rate. Participation matters. If you are not officially unemployed and you are not officially employed you are a missing participant.
January normally means that non-seasonally adjusted jobs drop. The only time that we did not see both full-time and part-time jobs drop during January was January of 2012 - probably people finding work in either the Obama campaign or the Romney Campaign. Will we see minor drops in jobs as we did during 2013, 2014, 2015 or will we see jobs fade as they did during 2006 or January 2010?
Normally we see seasonally adjusted jobs improve during January. It would be great for real workers if these jobs were real jobs instead of seasonally adjusted jobs. This number is rarely ever mentioned. What would the stock market do if it was reported that we added half a million SA CPS jobs as we did during January of 2016. They would either laugh or party like it was 1999.
We should see non-seasonally adjusted and seasonally adjusted unemployment levels increase - This is called January. We have already seen the Mid-January continuing claims data pop higher. It was also lower than it was during January of 2017. The continuing claims data is 100,000 claims lower than the same week of January 2017. The covered insured is only a fraction of the total unemployment picture. The official U-3 unemployment level. The unemployment level could pop up by a factor of four or more than the continuing claims data. Could the U-3 unemployment number jump "only 650,000" or less? It is possible . If the NSA unemployment level doesn't jump "enough" the seasonally adjusted U-3 value may negate some or all of the SA CPS growth. How will this be reported? If the seasonally adjusted jobs number increases and the seasonally adjusted unemployment level increases then the participation rate should increase, contingent upon the change in the population. Last year's downward revision to the population allowed the participation rate to rise instead of fall.
Participation Matters. The non-seasonally adjusted participation rate has been rising. The problem is that it was declining from 65.87% during January 2007 to 62.34% during January 2016, and it could have been recorded lower last year. The workforce population is revised during January. It can be revised higher, lowering the participation rate, or it can be lowered, thus raising the participation rate. The participation rate, NSA, normally bottoms during January. Currently the NSA participation rate is 62.69% This is well above the 62.45% rate during January 2017. The NSA rate could fall, especially if the number of unemployed workers doesn't spike and especially if the jobs numbers fall by a "normal" amount. Last year was a good year. This year could be even a better year. The seasonal factors are all over the place. Seasonal factors for the three components of participation cloud the picture. Non-seasonally adjusted participation should bump higher. Seasonally adjusted participation could bump higher.
The jobs article will discuss the U-7 unemployment rate that includes the non-participants and "effectively unemployed." The U-7 value has been dropping as more people rejoin the workforce. There will also be articles on how the multiple job holder data changes, the sector data, data for men and women and by age group.
It's the economy.
Copyright © Jack Dunn All rights reserved.