If you have "no place to go," come here!

BLS Jobs Report Covering January 2013: Terrible Month, Great Report

The short form: Yearly revisions increased the number of jobs created in 2012 by 647,000. These revisions make some comparisons difficult between December 2012 and January 2013 and obscure that January is an absolutely dreadful month for jobs and employment in real terms. After Christmas, the economy sheds large numbers of jobs that are not picked back up until later in the spring. The result is that while the adjusted numbers show gains, these numbers mark a trend basically bridging a chasm. The bottom of that chasm is where the economy now is. The number of employed dropped by nearly 1.5 million in the Household survey. The larger business survey documented just over 2.8 million jobs lost in January. These losses resulted in a real unemployment rate of 13.6% (versus a real trend rate of 12.6%) whereas real un- and under employment hit 18.9% (versus 17.4% trend) affecting nearly 31 million Americans. Blue collar workers lost ground in January in wages and hours, and in 2012 they also lost ground to inflation, low as it was. Average earnings for all private employees in 2012 increased about 2%, but this was probably all or nearly all wiped out by inflation. The CPI will come out later in the month. With that to the report.

The January jobs report is perhaps the most complicated of the year. There are year end revisions to both of the report's two surveys: the business or Establishment survey covering jobs and the Household survey covering employment (people). The business or jobs survey updates its numbers based on a much more complete tally for the month of March using unemployment insurance records. Tables for both seasonally adjusted and unadjusted data are revised. Revisions for unadjusted data go back to April 2011 and adjusted to January 2008. Based on changes in the March benchmark data, seasonally adjusted total nonfarm jobs in 2012 (the official jobs number) increased by 647,000 more than was previously reported: 134.021 million > 134.668 million

In the Household or people survey, revisions come from updates in Census data. However, unlike the business survey, these revisions are not applied retroactively. In other words, data for each year is treated as a done deal and all revisions between years are loaded into the January numbers. As a result, direct comparisons from the tables between Decembers and Januarys are not possible. However, the BLS does provide a supplemental table showing some changes taking these revisions into account for seasonally adjusted data only, and I will make use of those.

Turning to this month's headline numbers, the official U-3 unemployment rate was reported as "essentially unchanged" at 7.9% although it edged up a tenth of a percent in January. The seasonally adjusted official number of jobs increased by 157,000. In the regular two month revisions (yes, we still have those),

December increased 41,000: 155,000 > 196,000

November increased a further 86,000: 146,000 > 161,000 > 247,000 or a 69% increase over the initial estimate.

Month to month the potential labor force as represented by the Civilian Noninstitutional Population over 16 (NIP) increased 175,000 to 244.663 million. This NIP has built into it a yearly revision increase of 138,000. The employment-population remained unchanged at 58.6% with and without revision. So a rough estimate of the increase in jobs .586(175,000) would be 103,000 jobs. Per the official seasonally adjusted number 157,000, the economy did better than this. However, as I have pointed out many times, the official number is not a real number, it is a point on a trend line. In January especially, as we shall shortly see, the difference between the trend and what actually happened in the economy can be enormous. The trend smooths over what is basically a chasm existing between large after Christmas job and employment losses and the spring seasonal rebound.

The current labor force seasonally adjusted increased month to month just 7,000 to 155.654 million. The unadjusted number was 154.794 million with no direct comparison to December possible with the data in the tables.

Seasonally adjusted those employed dropped month to month 110,000 to 143.322 million. Unadjusted, the number was 141.614 million. While no direct comparison can be made with December, it is important to realize that this is 1.446 million less than December's 143.060 million. What is being masked here by the year end revisions is a huge seasonal drop in employment after the end of the Christmas shopping season. In real (seasonally unadjusted) terms, January is a terrible month, always.

So while the employment-population ratio seasonally adjusted remained unchanged. Unadjusted it dropped a staggering 0.6%. Remember each tenth of a percent is equivalent to around 245,000 people.

We see a similar effect in the unemployment numbers. Seasonally adjusted unemployment rose month to month by 117,000 to 12.332 million. Unadjusted, it increased 1.337 million to 13.181 million.

The unemployment rate seasonally adjusted, that is in official terms, increased a tenth of a percent to 7.9%. Unadjusted, it spiked from 7.6% to 8.5%.

The BLS broader measure of un- and under employment, the U-6, shows a similar discrepancy between the adjusted and unadjusted numbers. Adjusted, it remained unchanged at 14.4%. Unadjusted, it increased a whole percent to 15.4%

The seasonally adjusted U-6 un- and under employed represent the 12.332 million in the U-3 official unemployment rate, 7.973 million involuntary part timers, and 2.443 million marginally attached (those who have looked for work in the last year but not in the last 4 weeks before the survey) or 22.748 million total.

The BLS has a restrictive, though internationally recognized, definition of unemployment: without a job but have looked for one in the last 4 weeks. The marginally attached are not counted as part of the labor force and their use in the U-6 is an indication that this is what the BLS considers its functional undercount to be.
The BLS also has a more extended category: Not in Labor Force, Want a Job Now (seasonally unadjusted). This was up 249,000 in January to 6.781 million and could also be taken as a measure of its undercount. The problem which I state each month at this point in my analysis is that this number does not reflect very well actual changes in the economy. So I have developed an alternative to it. In my alternate calculation, I compare the current labor force to where we would expect it to be in a solid economic expansion: labor participation rate of 67% (boomers retiring so far has had little effect upon this figure). The difference between these two is my measure of the undercount. In addition, beginning with this post, I will calculate both trend (seasonally adjusted) and current (seasonally unadjusted) figures for the undercount and numbers derived from it.

.67(244.663 million) = 163.924 million (where the labor force should be)

Trend Undercount:
163.924 million — 155.654 million = 8.270 million

Current Undercount:
163.924 million — 154.794 million = 9.130 million

These are my estimates for the capture of the undercount, those who do not have jobs but would work if jobs were available to them.
With these numbers we can now calculate where the U-3 unemployment and U-6 un- and under employment (disemployment) really are, that is real unemployment and real disemployment and their associated rates both in terms of their trend and where they are now.

Real Trend Unemployment (that is seasonally adjusted):

12.332 million (U-3 unemployment) + 8.270 million (undercount) = 20.602 million
20.602 million / 163.924 million = 12.6%

Real Unemployment Now (i.e. seasonally unadjusted):

13.181 million (U-3 unemployment) + 9.130 million (undercount) = 22.311 million
22.311 million / 163.924 million = 13.6%


Real Trend Disemployment:

Real Trend Unemployment + involuntary part time workers seasonally adjusted = 20.602 million + 7.973 million = 28.575 million
28.575 million / 163.924 = 17.4%

Real Disemployment Now:
Real Trend Unemployment + involuntary part time workers seasonally unadjusted = 22.311 million + 8.628 million = 30.939 million
28.575 million / 163.924 = 18.9%

Looked at this way real unemployment (SA, seasonally adjusted) is 4.7% higher than the official (trend) rate while the real (NSA, unadjusted) rate for where we actually were in January 2013 is 5.1% higher. As I have often said, one of the easiest ways for officialdom to deal with unemployment is to define away as much of it as possible. Out of sight, out of mind, therefore not their responsibility.
The long term unemployed, those unemployed by the BLS definition, for 6 months or more was 4.708 million in January. They numbered 5.522 million a year ago. This is a decline of 514,000 but it is unclear where they went: into employment or out of the labor force altogether.
By race, unemployment among whites was little changed at 7.0%. Unemployment among African Americans declined slightly to 13.8%. The rough rule that unemployment among African Americans is twice that among whites continues to hold.
In the business survey, jobs (SA) increased 157,000 to 134.825 million. This represented 166,000 jobs created in the private sector and 9,000 job losses in government. Not seasonally adjusted (NSA), that is in the here and now, the economy lost 2.840 million jobs, dropping to 132.705 million, with 2.345 million lost in the private sector and 495,000 in government (almost entirely at the state and local levels). Among the big losers in the private sector was 90,000 in manufacturing, 272,000 in construction, and 1.974 million in the private services sector, including 592,000 in retail trade, 410,000 in professional and business services and 334,000 in leisure and hospitality.

Contrast this with the seasonally adjusted (SA) figures of gains of 33,000 in retail trade, 28,000 in construction, and little to no changes in manufacturing, professional and business services, and leisure and hospitality.

The total jobs numbers, the category called total nonfarm jobs, illustrates most clearly the difference between the trendline of seasonal adjustment and the actual swings of the unadjusted numbers which show where the economy actually is in any given month. In the following graph, the blue line is the seasonally adjusted trendline, and the red line is the seasonally unadjusted reality.

Seasonally adjusted versus unadjusted jobs

Average weekly hours for all private non-farm employees remained unchanged at 34.4 hours. Their average hourly earnings were $23.78, an increase of 4 cents, and their average weekly earnings were $818.03, an increase of $1.37 month over month, and 1.9% increase from last January.

Average weekly hours for production and nonsupervisory (blue collar) employees decreased a tenth of an hour to 33.6 hours. Their average hourly earnings were $19.97, an increase of 5 cents, and their average weekly earnings were $670.99, a decrease of 31 cents month over month, and a 1.2% increase over January 2012.

Household data (Employment/unemployment)
Statistical significance: +/ - 400,000
The A tables:
A 1 for most information and categories
A 2 Unemployment by race
A 8 Part time workers
A 12 Duration of unemployment
A 15 U 6 un- and under employment
A 16 Persons not in labor force

Establishment date (jobs)
Statistical significance: +/ - 100,000
The B tables:
B 1 Total jobs and jobs by industry/type
B 2 Weekly hours, all employees
B 3 Hourly and weekly earnings, all employees
B 6 Weekly hours, blue collar
B 7 Hourly and weekly earnings, blue collar

Average: 5 (1 vote)


Submitted by Hugh on

The one thing I have learned looking at economic data is that none of the major official indicators, like the job stats, GDP, or stock and commodity market levels says anything very useful about the actual state of the economy, the country, or us. At best, they are jumping off points for questions and analysis. At worst, they are obfuscatory.

I suppose the MSM is just a propaganda apparatus. They look at the numbers and parrot them back to us. It amazes me that they never bother to look further or dig deeper into them. The media seem filled with the most incurious people on the face of the earth. If I had their resources, and I wouldn't mind their pay either, I would be looking at this stuff simply to satisfy my own desire to understand. I would want to know why the official stats and the reality all around me didn't seem to be on even speaking terms.

Re the seasonally unadjusted data, people also seem to like nice lines and the seasonally adjusted jobs data (total nonfarm) has the nicest line around. Most of the rest of the adjusted data is less smooth and sometimes not all that much better than the unadjusted data. But isn't it interesting that the media and the government favor the smoother line, even if it is only marginally smoother and mistake it, deliberately or no, for the reality? And at the same time, they ignore the messier unadjusted data which is messier precisely because it reflects the messiness of real life.

I could easily make or probably I already have graphs at my FRED file that superimpose adjusted and unadjusted numbers for jobs and employment. But because of my slow connection I can't load them into the post. Oh well.

Submitted by Hugh on

It took me two hours to add that graph and it's still not quite the right size. Every time I go into an edit or preview page, it activates a slow running script. Anytime the screen changes or refreshes, that slow running script is activated. And each time it takes 10 minutes with my slow connection for the message to come up and almost that long for it to clear. I would use the edit buttons but they don't show up or at least not until I load a url into the transfer space, then they do, and then I hit the insert picture button, and got the url embedded in the text. But then it took me 3 tries and about 40 minutes for the preview page to load sufficiently for the save button to show up. Once I hit that the post loaded and displayed quickly and with no problems. I have the same problems with just text, but text is one time and its done so it only takes about 20 minutes to come up and 20 minutes to load and it's done. Adding one picture multiplies that time several fold.