Current Employment Commentary:
The U.S. Bureau of Labor Statistics (BLS) has released its “preliminary estimates” for the current U.S. employment situation for the month of April 2016. The numbers are considered “preliminary” for two more months to adjust for late arriving data and may be
adjusted outside these guidelines once again each January.
The number of Unadjusted jobs reported for April was 143.944 million up from January 2016’s 141.150 million. Employment is still below November 2015’s previous peak of 144.122 million.
Every January the BLS adjusts the previous data, often finding jobs that were “overlooked” previously. We have been tracking the data as released along with the changes since 2012. In 2013 they found 738,000 previously overlooked jobs, in January 2014 they found another 513,000 jobs and in January 2015 they found another 227,000 jobs. In January 2016 they backtracked and erased 75,000 jobs.
Also they say they adjust the numbers only a few years back. But when we actually verify their numbers pre-adjustment and post-adjustment, the changes often go much further back than they claim. The BLS calls this adjustment “Benchmarking”. This year the BLS “Benchmarked” their data back to 2000. Previous revisions have resulted in roughly a 1.5 million job difference, but the current revision resulted in very minor differences and even reduced the peaks in recent years minimally. They claim the most recent changes resulted from a new database for Services for the elderly and persons with disabilities coming on line which allowed the BLS access to more accurate numbers.
See BLS Changes Employment Numbers for the details.
On the Chart below we can see how the various sets of numbers compare. The 2016 revision is so close to 2015 that it is difficult to see. In addition to the issue of the numbers changing there is the problem that the “Civilian Non-Institutional Population” continues to increase by roughly 200,000 people a month, although from December 2014 – January 2015 there was a massive 700,000 increase… possibly due to the changes allowing illegal aliens to become legal at that time. The bottom-line though is that the number of jobs needs to increase more than the population increase for the unemployment rate to actually fall.
The Effects of Population Increases
According to the BLS, in November 2007 there were 139,443,000 people employed, the unadjusted unemployment rate was 4.5% and the civilian non-institutional population was approximately 232 million. In April of 2016 there were 143,944,000 employed for a net increase of 4.501 million jobs with an unadjusted unemployment rate of 4.7%.
But now there are 252.969 million civilians (not in jail, the military or institutions). So the population has increased by roughly 21 million.
The Numbers Don’t Add Up
Since almost half of the population is looking for jobs (Gallup says the payroll to population rate is roughly 45%) we would need roughly 9.45 million (21 x 0.45) more jobs just to be at the same employment level we were in 2007. But employment only increased by 4.501 million. So we are about 4.95 million jobs short. But somehow the unemployment rate is roughly the same???
According to Shadowstats the government is really underestimating unemployment by even more than our numbers suggest since “long-term discouraged workers were defined out of official existence in 1994.” The new U-6 numbers only include short-term discouraged workers.
The ShadowStats Alternate Unemployment Rate
Employment peaked in November 2007 at around 139 million and then fell to 136 million bounced above 138 million before falling below 128 million in January 2010. By December 2012 the number of people employed reached between 135 and 136 million depending on which BLS numbers you believe. A year later, December 2013, the BLS was saying the employment was above 138 million and by December 2014 the official numbers had topped 141 million.
From the chart, we can see that employment in November 2013 was nearing levels reached in November 2007 when employment peaked at just over 139 million jobs. But then it fell sharply as it seems to do every January (which is why they “seasonally adjust the numbers” to make annual comparison easier). Once again we need to note that the total U.S. population (not just the “civilian non-institutional population) has increased from about 305 million to about 323.5 million since 2007.
BLS Changes Employment Numbers
Note: In January 2016 the “non-adjusted” numbers underwent their annual massage (i.e. adjustment) and this time rather than creating phantom jobs they actually removed 75,000 phantom jobs.
But in January 2015 the “non-adjusted” numbers were adjusted and the BLS found roughly 227,000 jobs.
In January 2014 they found roughly 513,000 jobs. And even though they say they only go back three months with their regular adjustment, their annual adjustment can go back years. This time they went back and created more jobs out of thin air going all the way back to 1978. Just like with the “Ministry of Truth” in George Orwell’s Classic book “1984” it becomes increasingly difficult to determine where employment really stands and to track any trends, if the “yardstick” for measurement is constantly changing.
According to the BLS, “On an annual basis, the establishment survey incorporates a benchmark revision that re-anchors estimates to nearly complete employment counts available from unemployment insurance tax records. The benchmark helps to control for sampling and modeling errors in the estimates. For more information on the annual benchmark revision, please visit www.bls.gov/web/empsit/cesbmart.htm.” But that doesn’t explain why they would change data going back several years.
They said the “benchmarking” for January 2014 only went back as far as January 2009 but in actuality they changed the numbers all the way back to April 1979 — 20 years further than they reported!
For 2015 they “only” changed the numbers as far back as April 2013 and added 227,000 jobs. But the newest change is on top of the change they made in January 2014, which added 513,000 jobs and changed the numbers all the way back to April 1979.
In 2016, the BLS “Benchmarked” their data back to 2000. Previous revisions have resulted in roughly a 1.5 million job difference, but this revision resulted in very minor differences and even reduced the peaks in recent years minimally. These changes supposedly resulted from a new database for Services for the elderly and persons with disabilities coming on line which allowed the BLS access to more accurate numbers.
In 2013, the BLS changed the employment numbers back to July of 1991. At that time, I contacted the Bureau of Labor Statistics and they curtly pointed me to a short hidden blurb on their website that supposedly explains why they did it. See BLS Changes Employment Numbers for our full discussion of this (pseudo) BLS explanation. Their justification was that they hadn’t properly counted all the “Property and Casualty insurers”. And somehow Property and Casualty insurers increased the total number of jobs by 104,000 in August 2010 and by 738,000 in December 2012. Boy there must have been a lot of Property and Casualty insurers hired during that couple of years of recession!
You can see the cumulative difference between the 2012 numbers and the new numbers on the above chart the total difference is now 1.478 million jobs.
Note: The Employment rate and the Unemployment rate are based on two entirely different surveys but theoretically they should be two sides of the same coin. If you look at the chart above it does look like the employment rate is climbing i.e. more people are getting jobs as the trend channel does seem to be up.
By looking at the employment rate we should be able to tell how many jobs there are in our economy, pure and simple. (As Detective Joe Friday in Dragnet would say, “Just the facts Ma’am”). Are there more jobs than last year? Good! Are there fewer jobs than last year? Not good. Simple as that! But things have gotten a bit more complex with the implementation of Obamacare. It has inadvertently created incentives for companies to switch from full time employees to part-time employees thus two people may be considered employed where there used to be only one… except there is no additional work being done… but the numbers look better.
We need to consider the population factor. If the number of jobs stays the same but the population increases drastically then the unemployment rate will rise even though the number of jobs stayed the same. So looking at the unemployment rate is also important. Unfortunately, it still doesn’t give us the full picture. If the U.S. population is growing (and it is) but the number of jobs only holds steady the number of people without jobs will increase. So in order for the real unemployment rate to stay the same the number of jobs has to increase by at least the same rate as the population. If the number of jobs increases at anything less than the rate at which the population is increasing, the unemployment rate will be increasing.
Gallup polls look at the newly renamed “Good Jobs Metric” (formerly Payroll to Population aka. P2P) and has found that the number is surprisingly level from July 2011 (higher is better).
The P2P looks only at those working over 30 hours a week. So, in other words, when taking population increases and full-time employment into consideration, no matter what the official Unemployment rate says the P2P says that the employment situation is slightly worse than July 2011 (higher is better).
Instead of the P2P the U.S. Bureau of Labor Statistics uses the Labor Force Participation Rate. If we look at the BLS’ Seasonally Adjusted Labor Force Participation Rate we see that is has fallen drastically since 2009. Prior to 2009 it held fairly steady at around 66% of the workforce being employed but it is now below 63%. In January 2015, the Labor Force Participation Rate was 62.9% it slowly worked its way down to 62.5% in October 2015 and November but has now rebounded a bit back up to 63.0% in March 2016.
Labor force participation is still significantly below 2006’s 66% and nearing levels not seen since the 1950’s when the majority of women weren’t in the official labor force (current levels were last seen in the 1970’s when the country was in a massive recession and oil prices were skyrocketing).
Tom Thomas, one of our readers astutely observed that if you want to compare U-3 numbers you have to do it in light of the Labor Force Participation Rate (LFPR). The U-3 Unemployment rate during the Bush administration averaged 5.2% with an LFPR around 65%. Since U-3 only measures those who are actively looking for a job, if the labor force declines (i.e. lots of people stop looking) the U-3 rate will appear better than it actually is. So, in order to compare two time periods you have to adjust for the the LFPR. And if you are thinking that the current decline in LFPR is actually a result of more seniors retiring early (it’s not) see Record Low LFPR which shows the LFPR by age group since 1999. The biggest declines occur in the younger ages while those above age 60 actually show a higher percentage are working. For instance in 1999 only 24% of those age 65-69 were working but 14 years later 31% of those age 65-69 were working. This is not good… it means that people of normal retirement age are not able to retire due to economic reasons.
So why has the unemployment rate fallen so much since 2010?
As people stop looking for jobs (discouraged workers) the Labor Force Participation Rate (LFPR) falls. As the LFPR falls because of the way that the unemployment rate is calculated even though the number of unemployed stays the same (or even grows) it appears that the unemployment rate is falling.
To determine the employment rate the U.S. government surveys 390,000 businesses nationwide every month. The raw number is what we use here and it is not seasonally adjusted. You would think this number would be considerably more reliable than the 60,000 households that they survey to obtain the unemployment rate. And I prefer it to any seasonally adjusted numbers. This survey is submitted by the businesses monthly based on company employment on the 12th of the month but for some reason the first release is considered preliminary to allow for late arrivals and it often changes significantly in later months (when the public is no longer watching). For instance the October 2012 number originally said 134,792,000 (prior to the election) up from 133,797,000 the previous month but immediately after the election the October numbers were adjusted down -90,000 to 134,702,000. So rather than more jobs appearing due to late arrivals they somehow overestimated the number of jobs by 90,000. Typically, the variation during the 3 month adjustment period is no mare than 20,000. But somehow during this pre-election month they were off by 90,000.
See Current Unemployment Rate for an explanation of how the government calculates the official Unemployment rate.
Source: U.S. Bureau of Labor Statistics- Current Employment Rate Data
For more information See:
- Article: Employment vs. Unemployment how do they compare?
- Historical Employment Data Chart
- Current Unemployment Rate Chart
- Current Employment vs Unemployment Chart Are they really “two sides of the same coin”?
- What is U-6 Unemployment?
- The Misery index measures inflation plus unemployment and is a good measure of the discomfort of the country’s population.
- More Unemployment and Employment Charts
- 10 Awesome Jobs You Can Do From Home