AUG: +151,000 jobs. Unemployment rate steady at 4.9%. AUG details here!.. Jobs since Obama took office?... Unemp. rate under Obama?

Friday, September 7, 2012

August 2012 Jobs Numbers Summary

August Jobs Reports & Summaries:

  • August Summary: Report shows a weak increase in both total jobs and private-sector jobs but unemployment edges down as people left the labor force.  (This is quite a contradiction from the ADP private jobs report which was released yesterday and which showed an increase of 201,000 private sector jobs.)   
  • Economy adds 96,000 jobs; fewer than predicted.  Private sector jobs increase by 103,000; government jobs continue to decline; this month by 7,000.  
  • Unemployment decreases slightly to 8.1% as fewer people report themselves as unemployed, but fewer people report themselves as employed.  368,000 people left the work force with 403,000 more people saying that they "want a job".  
  • Manufacturing jobs decrease by 15,000; private service-producing jobs increase by 119,000.   
  • Unemployment rates for all ethnic and age groups declined except for an increase of 0.8% among young people 16-19 years old.  Unemployment for this age group increased 1.3% for both white youths and  black youths. The unemployment rate for Hispanic youth, however, declined 0.4%. 
  • Number of unemployed decreased by 250,000.  Number of employed decreased by 119,000.  
  • Employment decreased the most among white men and among those 16 -19 years of age of all ethnic backgrounds.  Employment increased among women.
  • Employment increased among people with Bachelor's degrees and some college, but decreased among people without high school degrees or with only high school degrees.  The number of people in the labor force decreased for both people without a high school degree and people with only a high school degree.  The number of people in the labor force with college degrees increased in August.
  • Employment dropped most among these groups:  Men and women 16 through 24 years of age.  Employment increased for men and women 25 years of age and older.
  • Employment increased by 43,000 for full-time workers and decreased 168,000 for part-time workers.
  • Number of self-employed workers decreased this month after months of increases. 
  • Numbers of unemployed people went up among people 16-19, 20-24, and men 25-34.  Numbers of unemployed people went down for all people 35 and over.  Unemployment went down the most among people 55+.         
  • Number of full-time workers increased by 43,000.  Number of part-time workers decreased by 168,000.   We have 1,982,000 more full-time workers than we did a year ago. 
  • What Was the Unemployment Rate When Obama Took Office Compared to Now? (August 2012)
  • Private, Government Jobs Gained, Lost Under Obama (August 2012 update)
  • Private Government Jobs Gained/Lost Month by Month 2011 & 2012 (August 2012 updates)
  • Jobs Lost/Created Year by Year Since 1999 (August 2012 Update)
  • How Many Jobs Have Been Created or Lost Under Obama?
  • How Many Jobs Have Been Created or Lost in August 2012?
  • How Many Jobs Created or Lost in 2012 To Date? (August 2012 update)


  1. When you say government jobs lost or gained (pretty much always "lost," recently), does that cover federal and state government jobs combined?

    My assumption is that it does, but I wanted to confirm.

    These are not just federal government jobs gained and lost, right?

    1. The losses have occurred mostly in the state and local government sectors, particularly in "local government - education". This month 10,000 jobs were lost between the state and local levels, but, over the past year, we've lost 83,000 jobs at the "local government - education" level, which is ... teachers.

      Federal government jobs have also decreased over the past year, a drop of 26,000 over the past year, not counting the post office.

    2. Thanks for the answer. So - just to be clear - if you're saying 10,000 jobs were lost at the state and local levels, that would mean 3,000 jobs were added at the federal level (as the overall loss in government jobs was 7,000)?

    3. Yes. 3,000 jobs were added in the federal government bucket. Unfortunately, the breakdown of those federal government jobs is not yet available.

      Since the beginning of the year, 23,000 jobs have been lost at the federal level, not including the post office. 6,000 of those were lost in the Department of Defense.

  2. One thing that is confusing people (and providing an opportunity for some people to intentionally mislead people) is what appears, at first glance, to be contradictory statements:

    * Economy adds 96,000 jobs.

    * Number of employed decreased by 119,00.

    These figures both come from the Bureau of Labor Statistics. But the first one comes from the BLS Current Employment Statistics survey, (CES), also known as the "payroll survey," whereas the second one comes from the BLS Current Population Survey (CPS), also known as the "household survey."

    For important information on how these two surveys are conducted, how employment and unemployment is defined in each (as well as some really useful graphs showing the longer-term trends) folks can go straight to the BLS website:

    In evaluating how much weight (if any) to give each of these figures, it is crucial to read the note on “Sampling error” on Page 5:

    “The payroll survey has a much larger sample size than the household survey. The payroll survey’s active sample covers approximately 486,000 business establishments of all sizes, representing about one-third of total nonfarm employment. The household survey is much smaller at 60,000 households, covering a very small fraction of total employed persons. Over-the-month changes in the household survey employment are therefore subject to larger sampling error, about four times that of the payroll survey on a monthly basis. When looking at short-term trends in either survey, especially over-the-month changes, it is therefore essential to assess the statistical significance of the change. (The sizes of the over-the-month changes in employment needed to be statistically significant are shown on page 4.)”

    Scrolling back up up to Page 4, you will find that the household survey employment numbers would have to have decreased (or increased) by at least 436,000 jobs in order to be statistically significant with a sample of the size that is used for that survey (in other words in order to conclude that the change was real, and not just statistical “noise”).

    The apparent loss of 119,000 jobs doesn’t even come close. In other words, this number, which many right-wing media sources and websites are trumpeting as "proof" that we lost jobs in August, is simply not a reliable number. Which is why that’s not the number that economists and (responsible) news organizations pay attention to. By contrast, the net gain of 96,000 jobs that the payroll survey showed IS above the number (90,900) that is required to be statistically significant in a sample of the size used for that survey.

    Bottom line: The figure of 119,000 jobs supposedly being lost is completely unreliable, from a statistical point of view, whereas the figure of 96,000 jobs supposedly being gained IS reliable, from a statistical point of view. Neither is likely to be totally accurate (and both will be revised later), but the +96,000 figure is likely to be much closer to the actual number.

    Yet right-wing media and websites are trumpeting the -119,000 figure as if it were a meaningful number. Is this because those citing the figures are ignorant and/or mathematically-challenged, or because they know full well that they are relying on a completely meaningless number but are doing so anyway because it fits their “Obama is wrecking the economy” narrative and they feel confident that their audience is biased and/or ignorant and/or mathematically-challenged enough that they can get away with it? Probably some of both.

    1. Yep. The CPS numbers are generally much more volatile, which is why the unemployment rate has gone up and down and the CES numbers have consistently gone up. The CPS (household) numbers also include agricultural workers and the self-employed AND it is basically self-reported. The CES number of jobs is consistently about 93.5% to 94% of the CPS number of employed. I also think the seasonal adjustments for the CPS numbers are a bit off. So we could actually have fewer people employed even if we added jobs, but I agree that it is most likely a sampling error.

      But the CES numbers are still off, as you mentioned, which is why we have benchmarking. The CES numbers are benchmarked in line with the actual quarterly unemployment tax reports that all employers are required to submit. A benchmark is scheduled to be published at the end of this month. For the past two years, the benchmark adjustments have added about 200,000 jobs each year.

      None of these monthly numbers were ever designed to be used as they are used, by stock market investors and political pundits. They are much more accurate and useful over time, and comparing from one year to the next is more revealing.

    2. Thanks, Molly.

      You have done a great job of explaining the various measures of employment and unemployment. Thorough, technically sound, and yet also highly readable. Plenty of attention to the big picture, as well as a detailed analysis of the most salient points.

      It would be great to see some of your pieces get a wider audience one or more of the high profile blogs. In the meantime, I will definitely be checking in here when the next round jobs reports comes out!

      - TRA


I appreciate intelligent comments and questions, including those that are at odds with anything posted here. I have elected not to screen comments before they are published; however, any comments that are in any way insulting, caustic, or intentionally inflammatory will be deleted without notice. Spam will also be immediately deleted.