Message Font: Serif | Sans-Serif

No. of Recommendations: 0

Being an Analysis by profession I got to wondering how Unemployment can go down if there are no new jobs created.  So I started down a path that is probably not worth going to see what the numbers tell me.  All the numbers come from the BLS release

Lets look at the U.S Employment data for October - November

Civilian Labor Force: Oct 153.975 mil ----> Nov 153.877 mil (net loss to labor force .098 mil)

Looks like 98,000 people left the labor force choosing to go back to school, join the military, retired, got incarcerated or decided to stop looking.

- Employed: Oct 138.275 mil ----> Nov 138.502 mil (net gain of .227 mil)

227,000 people who did not have a job now have one.  Must of been a good month for job growth?

- UnEmployed: Oct 15.700 mil ----> Nov 15.375 mil (net loss of .325 mil)

Unemployment is down 325,000 but only 227,000 people got jobs.  The difference? that equals the 98,000 people who left the civilian labor force and are no longer factored into the official unemployment number.

Not in Labor Force: Oct 82.575 mil ---> 82.866 (net gain of .291 mil)

Interesting only 98,000 people left the labor force but this number increased by 291,000.  This must be explained by 193,000 new people being included this month that just became eligible to be counted.  Some of these people could be newly turned 16 year olds, people leaving active duty military service, people returning to the workforce after being in retirement getting out of prison etc.,

Basic Math check: The sum of the Employed and Unemployed = the Civilian Labor Force

OCT  138.275 + 15.7 = 153.975   ///  NOV 138.502+15.375 = 153.877

Check Sat!

What does not in the Not in Labor Force stat Really Mean?

According to the BLS website: The Not in Labor Force stat includes everyone 16 years or older of those people who have no job and are not looking for one—" Many who are not in the labor force are going to school or are retired. Family responsibilities keep others out of the labor force.  Active Duty military are also not counted as in the Labor Force.  This is determined by asking the survey households some questions.

1.  Do you currently want a job, either full or part time?

2.  What is the main reason you were not looking for work during the LAST 4 WEEKS?

3.  Did you look for work at any time during the last 12 months?

4.  LAST WEEK, could you have started a job if one had been offered?

Now lets look at Jobs Data! Found in same Report.

Non Farm Employment: Oct 131.007 mil /// Nov 130.996 (net loss of .011 mil jobs) thats a loss of over 11,000 jobs.  That should have made employment more difficult maybe there are another source of jobs?

Since, Non Farm Employment jobs is the only statistic that the BLS reports we have to do some math on our own.  Should be easy though since we know the total number of people employed we can simply subtract the people working in Non Farm Employment, that should leave the total number of people employed in Farm Related work (or whatever else is not included in non farm equipment category)

Farm Employment (Total Employed - Nonfarm Job values)  Oct 138.275 - 131.007 = 7.268 mil  ----> Nov 138.502 - 130.996 =  7.506 mil (net gain .238 mil) Must of been a good harvest farm related employment rose 238,000.

Now lets look at the change in Unemployment numbers with the change in Jobs data

Civilian Labor Force: -.098 mil

- Employed: + .227 mil

- UnEmployed: - .325 mil

Not in Labor Force: +.291 mil

Total People in Survey (Civilian Labor + Not in Labor): +.193 mil

Non Farm Employment: -.011 mil

Farm Employment: +.238 mil

So according to the BLS data there were 193,000 more people accounted for this month compared to last month.  Since the difference between the sum of the change in Employed and Unemployed = the change in the Labor force we know that only 98,000 people left the work force to join the Not in Labor Force Statistic.  If you add the change in the civilian labor force 98,000 and add it to the change in total number of people surveyed 193,000 it is equal to the 291,000 people change in the Not in Labor force statistic (.193+.098 = .291).

This would mean that all of the new 193,000 people accounted for in the survey since last October are being reported in the Not in Labor force Statistic.  This effectively means that not a single person who became eligible to be counted in the Civilian Labor Force decided to seek employment.  So all 193,000 new people are either still in school, joined the military went to jail or Are Not Looking for Employment, or replaced someone in the workforce who gave up.  I think this is a little suspect.

Discounting that fact, we still need to account for the 227,000 previously unemployed people who found jobs that weren't employed before?  We know that the Non Farm Employment Segment shed 11,000 jobs so that means the Farm Sector picked up a staggering 238,000 jobs in one month.

I guess November is Harvest Season so its possible that the farming sector created 238,000 jobs.  Although the numbers I quoted are seasonally adjusted which according to the BLS defenition should have removed a large increase in farm workers normally attributed to climactic events like harvest season.

From BLS:

Over the course of a year, the size of the nation's labor force and the
levels of employment and unemployment undergo sharp fluctuations due to
such seasonal events as changes in weather, reduced or expanded production,
harvests, major holidays, and the opening and closing of schools.  The ef-
fect of such seasonal  variation can  be  very large; seasonal fluctua-
tions may account for as much as 95 percent of the month-to-month changes
in unemployment.

Because these seasonal events follow a more or less regular pattern
each year, their influence on statistical trends can be eliminated by ad-
justing the statistics from month to month.  These adjustments make non-
seasonal developments, such as declines in economic activity or increases
in the participation of women in the labor force, easier to spot.  For
example, the large number of youth entering the labor force each June is
likely to obscure any other changes that have taken place relative to May,
making it difficult to determine if the level of economic activity has risen
or declined.  However, because the effect of students finishing school in
previous years is known, the statistics for the current year can be adjusted
to allow for a comparable change.  Insofar as the seasonal adjustment is made
correctly, the adjusted figure provides a more useful tool with which to ana-
lyze changes in economic act.

Guess I will still need help from other people to tell me how you can lose 11,000 jobs and still increase employment by 227,000

No. of Recommendations: 0

It's the new math!

No. of Recommendations: 0

Good Question.  Nicely done, even I was able to follow the math.

No. of Recommendations: 0

the deception is gross here

No. of Recommendations: 0

You mean this wasn't evidence of yesterdays "Jobs Summit" working??

The BLS is just following in the steps of the Nobel Committee...  Providing recognition in advance of the accomplishments.

No. of Recommendations: 0

One of the problems with the data, is that unemployment statistics are based upon polls.  Unfortunately standard statistical analysis of polling data consistantly underestimates the margin of error of the poll, in my opinion.

There is no way of knowing how much of the unexplained difference is the result of the real margin of error for the survey.  With the number of employees being more than 100 million, any variations in the data less than 1 million, would be less than a 1% margin of error.  My intuition says that unemployment data likely has more than a 1% margin of error.

No. of Recommendations: 0

To count the number of people "working", I wonder if it would be easier to look at quarterly tax receipts and count the number of people who paid into the system?

Wouldn't we catch the people who are either salary, hourly or independent/small business?

Deriving these numbers from a survey seems like a very difficult task.

No. of Recommendations: 0

Here is the BLS defenition from http://www.bls.gov/cps/cps_htgm.htm

Where do the statistics come from?

Early each month, the Bureau of Labor Statistics (BLS) of the U.S. Department of Labor announces the total number of emplyed and unemployed in the United States for the previous month, along with many characteristics of such persons. These figures, particularly the unemployment rate—which tells you the percent of the labor force that is unemployed—receive wide coverage in the media.

Some people think that to get these figures on unemployment, the Government uses the number of persons filing claims for unemployment insurance (UI) benefits under State or Federal Government programs. But some people are still jobless when their benefits run out, and many more are not eligible at all or delay or never apply for benefits. So, quite clearly, UI information cannot be used as a source for complete information on the number of unemployed.

Other people think that the Government counts every unemployed person each month. To do this, every home in the country would have to be contacted—just as in the population census every 10 years. This procedure would cost way too much and take far too long. Besides, people would soon grow tired of having a census taker come to their homes every month, year after year, to ask about job-related activities.

Because unemployment insurance records relate only to persons who have applied for such benefits, and since it is impractical to actually count every unemployed person each month, the Government conducts a monthly sample survey called the Current Population Survey (CPS) to measure the extent of unemployment in the country. The CPS has been conducted in the United States every month since 1940, when it began as a Work Projects Administration project. It has been expanded and modified several times since then. For instance, beginning in 1994, the CPS estimates reflect the results of a major redesign of the survey. (For more information on the CPS redesign, see Chapter 1, "Labor Force Data Derived from the Current Population Survey," in the BLS Handbook of Methods.)

There are about 60,000 households in the sample for this survey. This translates into approximately 110,000 individuals, a large sample compared to public opinion surveys which usually cover fewer than 2,000 people. The CPS sample is selected so as to be representative of the entire population of the United States. In order to select the sample, all of the counties and county-equivalent cities in the country first are grouped into 2,025 geographic areas (sampling units). The Census Bureau then designs and selects a sample consisting of 824 of these geographic areas to represent each State and the District of Columbia. The sample is a State-based design and reflects urban and rural areas, different types of industrial and farming areas, and the major geographic divisions of each State. (For a detailed explanation of CPS sampling methodology, see Chapter 1, of the BLS Handbook of Methods.)

Every month, one-fourth of the households in the sample are changed, so that no household is interviewed more than 4 consecutive months. This practice avoids placing too heavy a burden on the households selected for the sample. After a household is interviewed for 4 consecutive months, it leaves the sample for 8 months, and then is again interviewed for the same 4 calendar months a year later, before leaving the sample for good. This procedure results in approximately 75 percent of the sample remaining the same from month to month and 50 percent from year to year.

Each month, 2,200 highly trained and experienced Census Bureau employees interview persons in the 60,000 sample households for information on the labor force activities (jobholding and jobseeking) or non-labor force status of the members of these households during the survey reference week (usually the week that includes the 12th of the month). At the time of the first enumeration of a household, the interviewer prepares a roster of the household members, including their personal characteristics (date of birth, sex, race, Hispanic ethnicity, marital status, educational attainment, veteran status, and so on) and their relationships to the person maintaining the household. This information, relating to all household members 15 years of age and over, is entered by the interviewers into laptop computers; at the end of each day's interviewing, the data collected are transmitted to the Census Bureau's central computer in Washington, D.C. (The labor force measures in the CPS pertain to individuals 16 years and over.) In addition, a portion of the sample is interviewed by phone through three central data collection facilities. (Prior to 1994, the interviews were conducted using a paper questionnaire that had to be mailed in by the interviewers each month.)

Each person is classified according to the activities he or she engaged in during the reference week. Then, the total numbers are "weighted," or adjusted to independent population estimates (based on updated decennial census results). The weighting takes into account the age, sex, race, Hispanic ethnicity, and State of residence of the person, so that these characteristics are reflected in the proper proportions in the final estimates.

A sample is not a total count, and the survey may not produce the same results that would be obtained from interviewing the entire population. But the chances are 90 out of 100 that the monthly estimate of unemployment from the sample is within about 290,000 of the figure obtainable from a total census. Since monthly unemployment totals have ranged between about 7 and 11 million in recent years, the possible error resulting from sampling is not large enough to distort the total unemployment picture.

Because these interviews are the basic source of data for total unemployment, information must be factual and correct. Respondents are never asked specifically if they are unemployed, nor are they given an opportunity to decide their own labor force status. Unless they already know how the Government defines unemployment, many of them may not be sure of their actual classification when the interview is completed.

Similarly, interviewers do not decide the respondents' labor force classification. They simply ask the questions in the prescribed way and record the answers. Based on information collected in the survey and definitions programmed into the computer, individuals are then classified as employed, unemployed, or not in the labor force.

All interviews must follow the same procedures to obtain comparable results. Because of the crucial role interviewers have in the household survey, a great amount of time and effort is spent maintaining the quality of their work. Interviewers are given intensive training, including classroom lectures, discussion, practice, observation, home-study materials, and on-the-job training. At least once a year, they attend day-long training and review sessions. Also, at least once a year, they are accompanied by a supervisor during a full day of interviewing to determine how well they carry out their assignments.

A selected number of households are reinterviewed each month to determine whether the information obtained in the first interview was correct. The information gained from these reinterviews is used to improve the entire training program.

No. of Recommendations: 0

Some numbers make me glaze over like a jelly doughnut. I am more interested in whether Russiangambit's friends have found work, or if oneLegged is finding more opportunities. Or if  Solaris and Tasty are needing to hire anyone because business is picking up.

No. of Recommendations: 0

"To count the number of people "working", I wonder if it would be easier to look at quarterly tax receipts and count the number of people who paid into the system?"

That would have the disadvantage that people doing the surveys would no longer be employed.

No. of Recommendations: 0

The BLS report is clearly separated into Household and Establishment data. You mix data from the two categories to arrive at your Farm sector numbers.

Labor Force, Employment, Unemployment, and Not In Labor Force data are categorized as Household data

Non-Farm Employment is Establishment Data

Household data is collected by Census Bureau interviews. Establishment Data is collected by BLS & State Agency questionnaires mailed to employers. Because the two data sets come from two different sources, the data is not interchangeable.

According to the BLS - "Data from these two sources differ from each other because of variations in definitions and coverage, source of information, methods of collection, and estimating procedures. Sampling variability and response errors are additional reasons for discrepancies. The major factors which have a differential effect on the levels and trends of the two data series are as follows." http://www.bls.gov/lau/lauhvse.htm

The numbers can only tell you what you let them tell you. No conspiracy here, but I gave you a rec anyway.

No. of Recommendations: 0

Well its interesting how the BLS describes their accuracy.

We have a 90% chance of being with 290,000 of a total survey.

Now does that mean 290,000 high of the actual number and 290,000 lower then the number?  Cause that would give you a range of 580,000.

Lets pretend that their number for this month is the number given from a complete census 15.375 million unemployed out of 153.877 million in the work force or 9.99% unemployed

the high estimate would be 15.375 + .290 or  15.665

the low estimate would be 15.375- .290 or 15.085

Assuming that their numbers for employed people are perfect and their is no error in their total sample size that would leave a range in unemployment between 9.80% and 10.18%

Although I seriously doubt that there total sample size is perfect so if your talking about a 600k range variance for both employment and unemployment numbers from the sample and your looking at some pretty huge swings.

Even still this basically says that comparing month to month data is stupid.

Also the government can't use tax receipts because that wouldn't count the employed illegal aliens.

No. of Recommendations: 0

Blake.. . I had meant it more as an exercise in why these numbers are pointless with people getting excited about a 0.2% decrease in unemployment like it meant something.

I clearly don't believe that there were 230k new farm jobs, after looking at the statistically significant description that they used in their own survey 230k is easily accounted for in their 90% probability of being within 290k of the number.

However that still leaves the question of why there was such a large increase in the total Not in Labor Force Statistic.

No. of Recommendations: 0

Thanks, i was hoping someone would do the analysis i was too lazy to do myself. I love this website.

No. of Recommendations: 0

The people that they called, had been unemployed so long their phones got disconnected.

They checked the box that said they stopped looking and moved on.

No. of Recommendations: 0

I clearly don't believe that there were 230k new farm jobs, after looking at the statistically significant description that they used in their own survey 230k is easily accounted for in their 90% probability of being within 290k of the number.

The 230k number is bogus because you were using inconsistent data. You should not believe it because it is wrong, not because of the accuracy of the survey.

However that still leaves the question of why there was such a large increase in the total Not in Labor Force Statistic.

Not really. If you have no faith in the accuracy of the unemployment data, you should have the same reservations about the Not In Labor Force data.

No. of Recommendations: 0

I like the ADP Employment Report vs. BLS data because it ties directly to payroll processing. Some paychecks may be severance checks, but the data represents real money being issued to real workers and is unaffected by the way someone in their HR department may answer a survey question. Seasonal and birth/death adjustments are not part of ADP results, unless of course, the worker was actually seasonally laid off or died. See how simple accurate data collection can be?

ADP data comes from 400,000 non-farm, private clients. This is a better indication of the state of U.S. employment because these private jobs represent the core source of taxes and profits that drive the economy. That's not to say a soldier or Social Security administrator is not important, but public jobs are ultimately supported by private enterprise (temporary {hopefully} excess printing of greenbacks, notwithstanding).

This week's ADP number showed a loss of 169,000 jobs in November and is one reason why the early Friday market euphoria quickly reversed. In my opinion, we already get a more accurate employment number from ADP for free and there is some tax savings opportunity at BLS.

No. of Recommendations: 0

Are we counting all those involved in the marijuana dispensary industry as being employed in the farm sector?  That may be where this number is coming from.

No. of Recommendations: 0

If the health care bill passes and the death rate increases, then the denominator (people available for employment) falls plus SCI starts hiring and the US enters a new death bubble industry with a falling unemployment rate.

To explain the current stats we have to include all the jobs that were saved according to President Obama, otherwise we would not have an accurate accounting of the improvement in the economy.

No. of Recommendations: 0

I was surprised that both September and October numbers were revised by 50%.  Don't they revise usually only for the prior month? Why are they revising September now?

And how can they be 50% wrong all of the sudden? The polls cannot be that wrong and if they afre, what good are they anyway?

I am not necessarily disputing the new numbers. They are as bogus as the old ones. But I am seeing jobs starting to come back here and there. I think we are leveling off to the new normal.