Because of the office move (broadband gets hooked up today tomorrow), I have been a few days behind in tearing apart major releases. Yesterday, we gave the business to last week’s Back-to-School sales data, which was weaker than it appeared.
We previously noted the many different ways BLS measures and reports unemployment (A Closer Look at Unemployment).
Today, we are going to identify an “issue” with the Unemployment Rate and Non Farm Payrolls — specifically, a look at the “official” U3 labor utilization rate, aka the 4.6% Unemployment Rate.
This Employment rate is expressed as a percentage. Its a simple fraction, with the number of employed as the numerator (the number on the top of the fraction) and the civilian labor pool as the denominator (bottom of the fraction).
This gives you a percentage; subtract that percentage from 100 and you get the Unemployment Rate. It looks something like this:
The Civilian labor force for August 2007 = 152,891; The Employed = 145,794. Put them in the formula above, and you get 0.95358130956 — approximately 95.4% Employment rate, and an Unemployment rate of 4.6%. The “Not Seasonally Adjusted” numbers generate approximately the same data: 0.953828513352
As we have seen in the past, there are two ways the Unemployment Rate can go lower, or stay. The “good” way is when more people get jobs. Then, the numerator goes or stays higher, and the percentage of unemployed goes lower.
The bad way is for the denominator to go lower. We have referred to this number in the past as NILF — Not-in-Labor-Force.
And that is precisely what happened to the Unemployment Rate in this month’s NFP. The BLS Household Survey (Table A-1. Employment status of the civilian population by sex and age) showed that in one month, the US labor force decline 1,379,000. (July 2007 = 154,871,000; August 2007 = 153,493,000).
That’s an awful lot of people who suddenly decided to “retire.” Similar numbers can be seen in the July 2007 “Not in labor force” — it ticked up from 77,087,000 to August 78,717,000). If it wasn’t for this sudden drop in the labor force, the Employment rate would have been = 0.945341606886.
In other words, a sudden and unexplained disappearance of a million plus workers from the labor pool managed to keep the Unemployment Rate almost a full percentage point below where it is more likely to be in that artificial construct called reality.
~~~
Some people will object to my not using “Seasonally Unadjusted” numbers. I have yet to figure out how these adjusting algorithms work. So let’s use the year-over-year numbers — no seasonal adjustment is necessary, as this compares an August to August. Persons “Not in labor force” declined increased from 76,702,000 in August 2006 to 78,717,000 in August 2007. That’s a drop of 2,015,000 over a year’s period. Some surely represent retirees, etc. But that’s an awful lot of people o have dropped out of the Labor Force in a single year.
The bottom line is that the seasonal adjustments and NILFs managed to artificially subtract about a percentage point off of the Unemployment Rate. Without that adjustment, the rate would be closer to 5.5%
~~~
Disclosure: We are short Korn Ferry (KFY), the exec recruitment firm.
I’m confused, how’s the disclosure about being short KFY related to the post?
~~~
BR Staffing agency. (My attitude is, its always better to disclose than not)
Is this supposed to be increased? Or persons in labor force declined?
~~~
BR: Doh! Damned double negative! Persons NOT in the labor force increased; The civilian labor force decreased.
I’ll fix it up top . . .
Come on Barry, don’t you see what is going on. W told us last year that the economy was strong and they were going to keep it strong. Fudging the calculations to get the data you want is one way to another mission accomplished. Should we expect anything else from Enron cronies? By the way we are also kicking ass in Iraq.
I seems to me that a rising NILF # for August when the economy as a whole is not creating piles jobs and sucking in illegals is not entirely surprising. Several million college students do go back to class in August. That is the sort of thing that seasonal adjustments attempt to smooth out.
If NILF rises in October it would be much stronger support for your arguement, you may well be right but August is a problematic month…
“Several million college students do go back to class in August. That is the sort of thing that seasonal adjustments attempt to smooth out”
Please note that BarryR addressed exactly that issue by also comparing across one year:
“So let’s use the year-over-year numbers — no seasonal adjustment is necessary, as this compares an August to August. Persons “Not in labor force” declined from 76,702,000 in August 2006 to 78,717,000 in August 2007. That’s a drop of 2,015,000 over a year’s period. Some surely represent retirees, etc. But that’s an awful lot of people o have dropped out of the Labor Force in a single year.”
Barry, a lot more people have dropped of the labour force presumably, what is scary is that 2 million net have dropped out of the labour force.
Ah but wait, it is all clear: those 2 million people are easily explained, they are just now living entirely off their tax cuts. Another splendid success for the Bush economic team! :-)
So Governments lie, so what else is new?
~~~
BR: Remember this quote:
“A Model is always wrong, but not useless.”
-Thomas Wilson, ING Group
The thing I have to keep hammering is that anyone who works for wages is a part of the labor force.
It easy to see labor force and think of the modal worker. A man, 30 – 55 who has a family, a house, etc.
However, most of the action happens to the marginal worker, who is 16 – 25, likely female, strong chance that she is a student.
When that worker can easily get a job waitressing, or as man busing tables, parking cars, perhaps even working in the mall he or she will.
However, when they have to compete against 30 year-old workers who are willing to work full-flex (i.e. any time the bossman calls) the younger workers will dropout.
Lowe’s is a great example. In boom times Lowe’s is full of kids, many of whom don’t know anything about home improvement. When the economy sours those kids are pushed out by craftsmen looking for work.
The kids were only working because Lowe’s pays what for them is a healthy salary. If they can’t work for that kind of money its not worth it to work.
Sadly I think that all those 50-somethings who are “retiring” will suddenly come back into the workforce when sadly they/their spouses need some $$$$ for unexpected health costs or other negative life event.
And don’t forgot the second largest (or is it largest? can’t remember) cohort of people in American history on the cusp of entering the labor market, the children of the boomers.
Talk about disinflationary labor pressures.
And the alchemy beat goes on. Criminals in well heeled suits!
OT
Anyone catch the daily show last night?? The piece on Craig (with the singer) was brilliant!! not to be missed…
BTW anyone who thought today (of all days) would even be allowed to end negative does’nt understand the first thing about psychology….
Ciao
MS
There is no doubt that statistics may be puzzling, governments may lie, BLS “birth/death black box adjustments” tend to under-/overestimate the job creation when business cycle turns … however, there was one interesting idea I saw in a research note by Stephen Lewis of Insinger de Beaufort (an anglo-dutch brokerage boutique):
“More likely, the contraction of 33k in jobs for temporary help in the three months to August is related to the raising of the federal minimum wage. The minimum wage rose from $5.15 per hour to $5.85 per hour on 24 July, the first increase since 1997. It would be surprising if this had no impact at all on the demand for labour. After all, in 1997 it was on 1 September that the minimum wage rose, a fact that may well have been associated with the freak fall in payrolls recorded in the August of that year … ”
OK, it does not solve “the big questions”, however – is there anyone with some kind of opinion about the possible impact from the “minimum wage adjustment”?
Nice post Barry. Just want one thing (good thing after you tore this unemployment report apart). On the year to year analysis, although 2 millions people suddenly became “not in the labor force” (78,717 – 76,702), we did add 1 million employed (employed went from 145,379 to 146,406.) For a net loss of 1 million in “reality”
Employment Population Ratio
Directly on point with our last post on NILFs and the labor participation rate, David points us to this delicious econ-wonk chart (below). It shows the relationship between a declining Employment Population Ratio, and subsequent recessions: Fascinating…
Good post Barry. For those who argue that workforce did legitimately shrink due to retirees, well if so, then it means the amount that has to get paid out related to unfunded liabilities is coming sooner than most think as the people are lining up faster than planned. Under estimate unemployment or under estimate fiscal strain…pick your poison but this box seems to be getting smaller and smaller and I really do not see a way out.
Barry: You should use the seasonally adjusted numbers. The workforce always falls in August and September because kids are going back to school. Just as it always rises in June.
The seasonally adjusted numbers show the same pattern anyway. Use them! It doesn’t mean you’re on the government payroll! The only problem comes when you try to compare seasonally adjusted employment against the birth-death model WHICH IS NOT SEASONALLY ADJUSTED.
A lot of the decline in the workforce comes from teenagers. We now have the smallest teen workforce since 1969. But the participation by adults is also falling.
~~~
BR: Agreed — I also used the raw year-over-year numbers also. they support the same concluson, as 2 million people out of the civilian employment pool since last August 2006.
I do not understand the mechanics of the seasonal adjustments.
I have a subjective measure of employment strength. The snottiness of employees in coffee shops. I tried to order a blueberry muffin in Manhattan Beach Sunday and the guy barked back “This is not Starbucks!”. Of course they hadn’t labeled their muffins so I had to take a guess.
Back when things were booming people seemed a lot more friendly and took pride in their jobs.
I object to your title and the phrase “artifically subtract”. This is the definition of the U3 unemployment rate. If you don’t like it, use another, but be sure to give historical numbers for comparison.
BTW, 1 million of those 2 million new NILFs are teenagers (ages 16-19). So it’s not just “retire”ment.
You make good points, Barry. No need to torture the data.
barry: the seasonal adjustment process is not that daunting. BLS uses a rather standard ARIMA model which compensates for variations from long run trend. They use the readily available X-11 (recently updated to X-12) model.
Barry, I see others are attempting to show how wonderful the numbers are, and what you should or should not use. I find it hard to argue with their point when I look at something as full proof as the Birth-Death Model which is unimpeachable. END SARCASM!
Just a personal note here. I recently took a retirement package from HP.
I’m 51, so I’m looking for job, but I haven’t applied for unemployment.
I wonder how my situation would be reconciled under the current
unemployment rate?
Mark Hessel
Barclay’s released a piece pointing to the JOLTS data as indicating private hiring has been solid and much better than the non-farm statistics. What are your thoughts on that? Thanks!
Just a personal note here. I recently took a retirement package from HP.
I’m 51, so I’m looking for job, but I haven’t applied for unemployment.
I wonder how my situation would be reconciled under the current
unemployment rate?
If you are available to start work now then you are unemployed. If you are not available then you are marginally attached but not in the labor force.
Whether you have applied for or even qualify for unemployment insurance is irrelevant.
BR: “Disclosure: We are short Korn Ferry (KFY), the exec recruitment firm.”
KFY is not a stock I follow so I pulled up a chart. Wow, nasty! If you were short before July (or even August), that was a hell of a nice call.
I’m curious as to whether the stock performance is company-specific or more of a macro thing. If the latter, it doesn’t bode well for the economic outlook. Could be an interesting tell since it’s not really a stock that the big boyz would prop up to make things look better than they are.
Could the presumed exodus(decline in remittance/wire transfers) of Mexican workers
explain the “retirements” from the workforce?
Well you also have all the people who aren’t unemployed but want work. The largest group being construction workers who are often sub contractors. I have many friends losing their butts because they can’t find work but don’t qualify for unemployment benefits. The other area is commissioned sales workers. REALTORS,mortgage brokers,car salesman. Not earning much if any money but not unemployed. That aside do you honestly believe anything the government says anymore?
OK, it does not solve “the big questions”, however – is there anyone with some kind of opinion about the possible impact from the “minimum wage adjustment”?
Let’s see. Service industries added 60000 jobs in August. Increases where in nurses, retail, food, leisure, hospitality etc. Financial services remained flat. Gov’t jobs dropped. Manufacturing dropped. Construction dropped. I don’t have the figures for IT services but I am guessing they either remained flat or dropped. Within leisure and hospitality, food services and drinking places employment continued to expand in August (+24,000). In professional and business services, management and technical consulting services added 7,000 jobs in August, and temporary help employment continued to trend down. Temporary help has lost 72,000 jobs thus far in 2007.
If it was the min wage you would’ve seen drops in food services, leisure etc but instead, they grew. Accounting for almost half of the jobs added in August. The drops where in the high paying jobs. Hope that helped.
http://www.bls.gov/news.release/empsit.nr0.htm
Correction on the above.
“Accounting for almost half of the jobs added in August.” should be “Accounting for almost half of the service jobs added in August.”
Karl Smith: “If you are available to start work now then you are unemployed.”
This is not correct. It’s that, plus “you have made specific efforts to find a job within the 4 weeks prior to the survey”. Specific efforts for example including contacting employers (i.e. pretty much interviewing or at least submitting resumes) but not just browsing job ads or networking your friends.
Age discrimination, often in the guise of “outdated skills” or “lack of passion for the job”, exists, and at some point the unemployable learn not to waste their time and efforts on jobs that are “not for them”.