Every month, at the beginning of the month the Bureau of Labor Statistics releases a bunch of numbers that allegedly describe the employment market that Americans find themselves in. These numbers and some others from other sources are then fed into the media spin machine and talking heads hyperventilate about what all this means for the election prospects of the candidates. Amidst all of the spinning, we are left to wonder, “do these numbers really mean anything?”

An old couch with a 'Jobs' sign

Photo: Doug Geisler / Flickr

It would be good to consider whether the numbers are meaningful in two ways; do they mean anything both in terms of whether they truly reflect the experience of Americans in the job market and does the growth in employment that they sometimes state represent significant and appropriate progress?

The headline number that gets reported every month by the Bureau of Labor Statistics (BLS) is called the U3 number which is calculated from two seperate surveys, the household survey, otherwise known as the current population survey and the establishment survey. The establishment survey is a sample of jobs as reported by employers and excludes agricultural workers. The household survey includes estimates of the self-employed, farm workers and even workers that don’t get paid, for example, relatives working without pay in a family business. (That’s right, you don’t have to get a paycheck or have a formal contract to be counted as employed.)

The establishment survey is based upon employer reports of the pay period which includes the 12th day of the month. For the establishment survey, if a day laborer isn’t paid on the 12th, then that isn’t counted. If a casually employed worker works one hour in a whole month and that hour is on the 12th, he’s considered employed. The household survey reports results for the week which includes the 12th day of each month. Workers are counted as employed by this survey even when they are absent from their jobs for that entire week, paid or not paid.

The margin of error for the household survey is about 436,000 and for the establishment survey is about 100,000.

The U3 report notably does not represent a number of categories of workers that are reasonably considered unemployed, such as those who have exhausted their unemployment benefits, nor does it represent the struggles of workers who find themselves employed in part-time jobs or, find themselves employed in work that does not reflect their training, education or abilities. These numbers say nothing about, for example, recent college graduates who have “played by the rules,” invested heavily in their education and find themselves heavily in irremediable debt to the same financial sector which crashed the economy and destroyed their job prospects.

Since many politicians, pundits, talking heads and bloggers are taking employment numbers as a bellwether for the President’s electoral fortunes, if their political analysis is to be correct then the best employment number to use would be the one that bears the closest relation to how the electorate experiences the employment market.

Considering that the (frequently revised) reports of jobs added or lost by the U3 are often within the margin of error for the surveys and the definition of what is counted as an employed person is a little odd, it makes one wonder if there isn’t a better measure of unemployment available. As it turns out, the BLS has alternative measures available, with U6 being the broadest measure that they offer. The broader U6 includes “people who want and are available for full-time work but have had to settle for a part-time schedule” and those that indicate that “they want and are available for a job and have looked for work sometime in the past 12 months.” The most recent U3 number is 8.3% while the U6 is a much higher 15%.

Gallup polls also does an employment survey with a somewhat different methodology than the BLS surveys. Gallup’s poll is not seasonally adjusted, surveys the population 18 and older (while the BLS covers age 16 and over) and the sample size is smaller than BLS’. Gallup’s number this month for its poll that is closest to U3 is 8.2% and the Gallup number for their poll that is closest to U6 is 17.1%.

Given what the various measures on offer include, it appears that the numbers that most closely reflect people’s experiences of the employment market are the U6 numbers or the Gallup survey numbers that count the underemployed, though those don’t reflect how many people are working in inappropriate jobs. Here’s a some BLS data in chart form an article published in the Chronicle of Higher Education that tracks college graduate workers in various non-professional occupations:


America seems to have a pretty impressively well-educated workforce when nearly 1 in 7 of our parking lot attendants have at least a bachelors degree. On the other hand, it seems more likely that the fact that so many well-educated people are “choosing” to work in occupations where they cannot realize the potential of their education and training is an indicator of low demand for labor.

Here’s an enlightening set of data from the BLS, the employment population ratio:

august emp-pop ratio 1

or here it is in chart form:

august emp - pop chart

What this chart shows is that the portion of the population that is employed peaked in 2007 at about 63% and suffered a precipitous drop to a little above 58% and has remained around a little above 58% oscillating up and down small amounts since dropping.

In light of this evidence, the headline news and the monthly trepidation about the U3 employment numbers seem to be much ado about nothing.

The employment picture is not improving significantly for American workers and it appears that despite all of the assorted programs put into place, the bailouts of banks and corporations to “save the economy,” the heroic exertions of politicians to cut the deficit, etc. – nothing that has been done to date has improved employment since it reached a “new normal” early in Obama’s presidency.

joe shikspack

joe shikspack

1 Comment