Bookmark and ShareShare
Friday, September 7, 2012

Don’t Be Misled by the U.S. Unemployment Rate

U.S. unemployment was 8.1% in August, according to the U.S. Bureau of Labor Statistics (BLS). Meaning, of all the people who want to work or are actively looking for a job, 8.1% don’t have one. This number gets an awful lot of media attention, particularly in an election year, but it is deeply misleading. For example, in August, relatively few jobs were added but more workers dropped out of the workforce, so the unemployment rate went down.

Most leaders and citizens probably don’t know what the unemployment number represents. I suspect most don’t even know that it’s actually a survey of 60,000 households conducted each month by the Current Population Survey on behalf of the BLS. Some think it is an actual tally of paperwork from claims. It is not -- it’s an estimate drawn from monthly tracking surveys.

The problem with the unemployment metric is that an additional 6.6% report being “underemployed,” meaning they are left out of the 8.1% unemployed because they have part-time work but wish they were employed full time; they’re not classified as “unemployed.” They’re not in the 8.1% everyone’s watching, and this makes things complicated. Nearly 15% underemployed, including the unemployed, is much more accurate and significant than the single 8.1%.

Things get even more complicated. Some examples:

  • If someone has lost all hope of finding a job and just stays at home deeply depressed, the government doesn’t count him as “unemployed,” because he is not actively looking. I’d assert that “hopelessly unemployed” seems as unemployed as one can get.
  • If a laid-off engineer comes over and mows your yard, taking one hour of time, and you pay him, say, $20 -- and that is the only hour he works that week -- the government doesn’t count him as unemployed, but as underemployed. So that laid-off engineer is officially not unemployed. How does that make sense?
  • When several hundred thousand workers get discouraged, the size of the U.S. workforce mathematically decreases, because fewer workers report “looking.” So even if jobs are decreasing, it will appear unemployment is improving because the denominator got smaller --driving the percent unemployed down. This is really misleading. 
  • And one more thing: The unemployment metric gets “seasonally adjusted,” meaning the BLS tries to mathematically correct big swings in people going in and out of the workplace by weighting the data according to things like warmer weather, vacations, back-to-school season, holidays, etc.
My big point here is that the current U.S. unemployment metric is an over-complicated mess and misleads smart American leaders and citizens. Bluntly, we don’t honestly know when we see the government’s unemployment data if the employment situation got a little better or a little worse.

Gallup has a solution and breakthrough. We will start reporting Payroll to Population employment rates (P2P) every day on our website as of today. We will use a big sample of 30,000 completed telephone interviews in the U.S. and compute a new global employment metric that will be pure and unadjusted.

Our first calculations are out, both for the U.S. and the entire world. When P2P moves, it means real unemployment numbers moved.


Anonymous said...
September 7, 2012 at 6:01 PM  

Have you given any thought to adjusting the P2P rate for demographics? As the Chicago Fed explained earlier this year, almost half the decline in the labor force participation rate over the past decade can be explained by demographics, "such as the retirement of the baby boomers," and the pattern is expected to continue (baby boomers began turning 65 last year). Accordingly, the Chicago Fed has recommended that measures like "the employment-to-population ratio" should be adjusted to "reflect these long-running patterns."

Unknown said...
September 7, 2012 at 7:50 PM  

Is this the way unemployment has always been calculated? If so, then while I certainly agree that the value is misleading, it seems to me a number of people have picked up on this as well), but fail to put the number in context and are simply stating the current number for shock value, rather than looking at how an adjusted unemployment calculation would have affected historic rates as well as current rates.

Milton said...
September 10, 2012 at 9:42 AM  

Kudos to Gallup for having the courage and backbone to report stats that are a true reflection of employment reality, despite public pressure from the "emperor" whose "recovery" appears to have no clothes at all.

Anonymous said...
September 10, 2012 at 6:49 PM  

As you know, for years the BLS has consistently published a variety of unemployment rates, some of which do count underemployment; in addition, they publish non-seasonally adjusted numbers. The extended data is not a secret. If "smart American leaders and citizens" are misled, it's because they don't bother to understand the primary data being reported in the news. I wish you the best of luck in communicating the nuances of your P2P rate to this audience.

Center PAC said...
September 11, 2012 at 11:49 PM  

During the next few years Medicare rolls will increase to 75 million (from 40 million in year 2000). What effect will these tens of millions of people have on your payroll-to-population ratio? I bet that we will get to 5% unemployment rate in only 2 years, yet your payroll-to-population ration will continue to drop.

Otherwise ... I agree. The monthly unemployment surveys are not reliable. They they do paint a picture over 6 months or more as long as they do not change methodology.

Post a Comment

Comments are moderated by Gallup and may not appear on this blog until they have been reviewed and deemed appropriate for posting.

Copyright © 2010 Gallup, Inc. All rights reserved. | Terms of Use | Privacy Statement