BLS Jobs Report, November 2011: Good News! But I Have Doubts
The Bureau of Labor Statistics jobs report covering November 2011 is out. The big numbers are an extraordinary drop in the official U-3 unemployment rate of 0.4% from 9% to 8.6% in the Household survey and a respectable jobs number, 120,000, from the Establishment survey. Additionally, September's jobs number was revised upward 52,000 from 158,000 to 210,000 and October's 20,000 from 80,000 to 100,000. If you have a good memory, you will recall the original jobs number for September was 103,000. Subsequent revisions have more than doubled it. You can look on this as further good news or as further evidence that BLS figures are becoming increasingly unreliable.
Diving into the data, the civilian non-institutional population (NIP), the potential labor force, increased by 172,000 from 240.269 million to 240.441. The employment-population ratio increased a tenth of a percent to 58.5%. Multiplying this ratio by the change in the NIP gives us 101,000, the number of jobs needed to keep up with population growth. The economy did a little bit better than that.
We begin to see what is really going on when we look at the employment/unemployment numbers. The number of unemployed dropped by a whopping 594,000 in November from 13.897 million to 13.303 million. Part of this decrease is accounted for by an increase of 278,000 in the number of employed which went from 140.302 million to 140.580 million. However, something over half of the big change in unemployment (315,000) came from the BLS' arbitrary policy of defining certain of the unemployed out of the labor force. The actual labor force decreased from 154.198 million to 153.883 with a corresponding decline in the labor force participation rate (the ratio of the actual labor force to the larger potential labor force as represented by the NIP) from 64.2% to 64.0%.
As you see, there are two parts to this story. One is the seasonally adjusted increase of 278,000 in the number of employed. That's good news, but when there is a big swing like this, it is always good to look at the unadjusted numbers to see what they add or subtract to the narrative. And here it is worth noting that seasonally unadjusted numbers show an increase of only 83,000 in the employed, or less than a third of the seasonally adjusted number.
The other part of the story is the labor force participation rate. During a solid, not mammoth, expansion, I take this rate to be around 67%. Multiplying this by the potential labor force of the NIP gives us an expected number of employed of 161.095 million. Subtracting the current labor force 153.883 million from it gives us my calculation of the BLS undercount or 7.212 million.
So this tendency of the BLS to arbitrarily send some out of the labor force adds up after a while. It also has other effects. For example, the BLS can report that the number of long term unemployed (27 weeks or longer) decreased by 185,000 last month. Whereas if we took the 315,000 the BLS removed from the labor force, it increased by a 130,000.
The BLS partially addresses this with its broader U-6 measure of un- and under employment. The U-6 is, you will remember, comprised of the U-3 official unemployment, part-timers who would work full time if they could, and the marginally attached. The marginally attached are something over a third of the BLS undercount and represent those who are not looking for work now but want a job and have looked for one in the last year. This rate also fell significantly in November from 16.2% to 15.6%. The fall came due to the big drop in unemployment and from a decrease in involuntary part timers of 378,000 from 8.896 million to 8.518 million. The number of marginally attached increased slightly (36,000) from 2.555 million to 2.591 million. Multiplying the U-6 rate by the labor force (153.883 million) gives us a figure of 24.006 million for the U-6.
Turning to disemployment, the figure which takes the BLS undercount fully into account, I use two different methods. In the first, I use the seasonally adjusted BLS category Not in Labor Force, Want a Job Now, as a BLS estimate of its undercount. My chief criticism of this number is that, while it gets us into the ballpark, it doesn't vary all that much even in the face of major economic events, like the December 2007 recession. In November, this number was 6.595 million, an increase of 192,000 over the previous month. Note the difference between this and the 315,000 which the BLS disappeared from the labor force.
Subtracting out the overlap of the marginally attached gives us 4.004 million. Adding this to the U-6's 24.006 million gives a disemployment number of 28.010 million and a disemployment rate of 17.7%. This is half a percent better than October and a full percent better than September.
For my own calculation of disemployment, I take my figure for the BLS undercount as I set it out above .67(NIP) or 7.212 million, (an increase of 430,000 from last month's 6.782 million) and add it to the number of involuntary part timers, 8.518 million, and the unemployed, 13.303 million. This shows the disemployed at 29.033 million, down 542,000 from October, still a large decrease. The disemployment rate decreased from last month's 18.4% to 18%. There has been a 0.7% decrease in disemployment since September.
These numbers suggest something is going on but what isn't clear. Is what we are seeing real or an artifact of the BLS' models? For these kinds of changes, you should be able to see corresponding improving signals over the last two months in your communities. The question is have you?
Traditionally, in an improving economy, we would expect to see both wages and hours to increase, but in November the average work week remained unchanged at 34.3 hours and average hourly wages decreased slightly, by 2 cents, to $23.18. Basically, Americans are still working fewer hours than they were in 2007 (94.3%) and though nominal wages have increased by 4.3% since then real wages (taking into account inflation) have been falling.
Other significant items in the report include that its good news fell along racial lines. White unemployment decreased from 8.0% to 7.6% while African American unemployment which had been falling over the last few months headed back up from 15.1% to 15.5%, and is once again double what it is for whites. The big winners were white males over 20 who saw their unemployment rate dive from 7.9% to 7.3%. White women, on the other hand, saw their unemployment rate nudge down only slightly from 7.0% to 6.9%. White teenagers (16 to 19) saw their unemployment decrease from 21.8% to 21.4% while African American teens saw theirs increase from 37.8% to 39.6%.
Turning briefly to the Establishment data, the economy has created net 337,000 jobs in the last 12 months. The private sector created 1.096 million jobs. Government lost 759,000. Healthcare which is often touted as a big job creator produced 292,200, but professional and business services created, 353,000, and leisure and hospitality produced even more 499,000. Retail trade lost 121,300. Manufacturing increased by 160,000 while construction decreased by 123,000. So what does all this say? That there are winners and losers on both sides of the pay divide. But overall more good jobs (government/construction vs. manufacturing) were lost and more poorer ones (leisure and hospitality vs. retail) were created.
What we are seeing, in other words, are fluctuations within the larger trend. In the last 12 months, the economy needed to create 1.010 million jobs just to keep up with population growth but it only produced 337,000 jobs. So we remain substantially down and what jobs are being created are not as good as the ones being lost. At the same time, according to the BLS, 430,000 jobs were created in the last 3 months alone. So we might expect signs of improvement. There are two important questions here. First, are the jobs real? And, second, if they are, is their poorer quality negating any positive economic value from them? There should be discernible improvement in the economy accompanying them but, looking out my window, I don't see it.