Jobs Number, Part 2: Digging In

by March 21, 2017
There's more than one unemployment figure each month. Here's how to dig into the report. (Image by "pixel2013" via pixabay CCO Public Domain)

There’s more than one unemployment figure each month. Here’s how to dig into the jobs number report. (Image by “pixel2013” via pixabay CCO Public Domain)

Proper reporting of the monthly employment report takes work. Read the BLS analysis. Along with notes and footnotes, it will explain factors which may have influenced the report. Report on the trends month to month and compare year to year. These trends offer much more insight than a single monthly number. They also make for nice graphs and charts.

Household survey basics

Be aware that the employment report is two separate surveys. For the household survey, field workers knock on doors and ask, “Are you employed?” If the answer is “no,” they will follow up with the question, “Are you looking for work?” This produces the widely reported unemployment rate, currently 4.7 percent. Be aware it actually produces six rates.

The rate most often used in headlines is known as U-3, the officially recognized rate of unemployment. A much broader rate, the U-6, tries to capture how many people want to work but have given up their search. This is the number many critics point to when they argue the unemployment rate undercounts the number of unemployed. In fact, the employment report has all this information in it. It is up to reporters and editors to decide what information to use in their stories.

There is a lot of “noise” in the household survey. Noise is a perfect acceptable term statisticians use for unexplained and unpredictable variation in the data. People tend to start and stop their job search. If you lose your job in November, you probably won’t look for a new one until the start of the next year. Of if you read that hiring is picking up you might start looking after having thrown in the towel.

Establishment report basics

The second survey, known as the establishment report, goes after employment data from the other direction. It asks companies how many people they have on their payroll. This produces the “added to payrolls” or “job creation” number, which was 235,000 in the February report. Companies have a pretty good handle on how many people they are paying, so this is considered the more reliable of the two surveys. The trend in this number is also easy to compare with a similar number reported independently by Automated Data Processing (ADP), a public company providing human resources and payroll services. Its monthly employment report monitors a subset of its payroll processing clients and provides a good check on the BLS data.

Because these are different surveys, they cannot be directly compared. It is not unusual for the two BLS surveys to point in opposite directions in a month. Or for the BLS establishment report to disagree with the ADP survey. But the trends should be, and are, consistent over time.

The participation rate

There is truth to the observation that the “participation” rate, that is, the number of people in the workforce, has been declining. It is not, however, a matter of fraud, as some critics imply. In part, it is due to baby boomers reaching retirement age. But a significant part, and one worth writing about, is automation. You will often see these show up in economic reports as increases in productivity.

For every job that left America because it could be done more cheaply by workers in another country there are several jobs which have been lost to automation. While changes in tax policy can bring some jobs back to the United States, many will be performed by automated equipment, not by humans.

Manipulating the numbers

Numbers are funny things. Even though they appear to be absolute, a clever manipulator can twist them to make pretty much any point he wants to make. Take Trump’s statement from February: “Ninety-four million Americans are out of the labor force.” It might seem preposterous but it is correct, as the great sage Obi-Wan-Kenobi once said, “from a certain point of view.”

It is the number you get if you take the total U.S. population 16-years of age and older and subtract the people the BLS says are in the labor force. The number includes everyone who is retired, and most high-school, college, graduate or vocational school student. It also includes the disabled, homemakers, some self-employed and those living off their investments.

Reporter’s Takeaway

• The BLS Employment Situation Report is actually two seperate surveys. The “Household” survey generates the unemployment rate. The “Establishment” survey reports changes in the number of people employed.

• Automation and other changes in productivity are the cause of much of the job loss in the United States.

• Reporters should be suspicious of claims about the validity of the numbers, but claims are often grounded on at least a modicum of truth.

In his longtime role as New York Bureau Chief and Senior Correspondent for public television’s Nightly Business Report, Scott Gurvey covered the financial markets, business and the economy for the first and longest-running daily program dedicated to financial news on broadcast television. Currently Gurvey writes for various online sites, teaches the next generation of journalists and advises companies and nonprofits on media relations, communications strategy, ethics, newsroom operations and online application design and practice. His blog, “Public Offering,” lives at, and Gurvey has been known to Tweet at @scottgurvey, from time to time.