Veteran business journalist” and Fox News talent Stuart Varney is choosing to place his reputation on the line over the jobs numbers that came out this morning. It’s a big story: The country’s official unemployment rate fell by three tenths of a percentage point, from 8.1 percent to 7.8 percent. Channeling unemployment-report conspiratorialist Jack Welch, Varney said, “Oh, how convenient that the rate dropped below eight percent for the first time in 43 months five weeks before an election.” Welch today tweeted, “Unbelievable jobs numbers..these Chicago guys will do anything..can’t debate so change numbers.”

More Varney:

There is widespread mistrust of this report and these numbers because there are clear contradictions---873,000 people said that they had found work. But only 114,000 new jobs were created. That is a contradiction.

On their face, the numbers appear to invite media conspiracy. Those 873,000 people who indicated they’d found work? That number is gigantic relative to historic monthly changes. The stats show nothing so drastic since 1983, according to Stacey Standish, a spokeswoman for the Bureau of Labor Statistics (BLS).

The way Varney and Welch talk about these numbers, it’s as if they’re just guesstimates that get tossed around in Washington. Just impressionistic gauges of employment.

Yet look at how BLS comes up with the figures. It relies on the Census Bureau to conduct a survey of 60,000 households. The process consists of a Census person quizzing people about whether the folks in their household have work. Given that the whole process is a survey and depends on information supplied by humans, it’s never perfect. But it is thorough. Here’s a passage from the Bureau of Labor Statistics website:

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....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.

A separate survey published by BLS focuses on payrolls. According to BLS spokesperson, approximately 400,000 “work sites” across the country are surveyed to determine the level of job creation in the economy. As noted by Varney, the latest report indicates an increase of 114,000 nonfarm jobs.

Now onto what Varney terms a “contradiction” in the two parts of the September jobs report: How, he asks, can the survey show that 873,000 people have flooded the workforce while only 114,000 new jobs have been created? Here’s how Varney puts things:

If you delve a little deeper, seems a lot of these people who found turns out that 600,000 of those 873,000 people were part-time workers, so they came back into the labor force and they pushed the unemployment rate down to 7.8 percent.

Herewith a textbook example of how the media botches unemployment numbers. In fact, Varney’s reference to the 873,000 bump being driven by a population of 600,000 part-timers misstates the complexity of the survey.

The 600,000 figure is a rounded-up version of the 582,000 uptick for September in the number of people employed part-time for what the BLS calls “economic reasons.” That means that these people are essentially non-voluntary part-timers and would prefer to be more fully employed. The overall number moved from roughly 8.0 million to 8.6 million.

However, that the number of people who reported being involuntarily employed part-time jumped by nearly 600,000 doesn’t mean that there were 600,000 new part-time workers entering the labor force. It means only that roughly 600,000 people reported being employed part-time involuntarily. A goodly yet indeterminate fraction of that group could well be people who were downgraded from full-time work to part-time work. “It’s not necessarily that people are coming into the labor force,” says BLS economist Vernon Brundage. “Their hours could have been reduced.”

Another complication: Not all part-timers are involuntary part-timers. Over the past month, the population of willful or even happy part-timers actually went down by 260 million to 18.7 million people. That reduction further undercuts Varney’s point about the swelling of the part-time economy.

Emy Sok, an economist at BLS, addresses the perils of pinning the 873,000 boost to a single dynamic in the labor force. “I don’t think you could point to any one group as the main driver,” says Sok, who does note that employment among young workers did increase “more than typically.”

Now to this “contradiction” in the two numbers addressed by Varney. How, indeed, can you have a boost in the employment level of 873,000 alongside an increase of just 114,000 nonfarm payroll positiions? As mentioned above, the jobs report sits on top of two separate surveys — the household survey that determines how many people are employed and the payroll survey on job creation. Sok explains that the household survey counts agricultural work, unincorporated self-employment, unpaid family workers, private household workers and unpaid absences. That’s informal economic activity that won’t be detected in the payroll survey. Which is to say that some unevenness between the two sets of data is almost baked into the process.

People like Varney — and many others out there today in political fantasyland — who are inveighing against the integrity of the jobs numbers would be well advised to spend some time inside this pdf. It’s called “Technical Paper 66,” a production of BLS and the Census Bureau. Chapter Two of “Technical Paper 66 ” delineates all of the changes that the government has made to procedures in the household survey, aka Current Population Survey (CPS). The capsule summaries of methodological updates confer a decent idea of how fastidious this enterprise is. Here are a couple of my favorites:

September 1975. State supplementary samples were introduced. An additional sample, consisting of about 14,000 interviews each month, was introduced in July 1975 to supplement the national sample in 26 states and the District of Columbia. In all, 165 new PSUs [Primary Sampling Units] were involved. The supplemental sample was added to meet a specific reliability standard for estimates of the annual average number of unemployed people for each state. In August 1976, an improved estimation procedure and modified reliability requirements led to the supplement PSUs being dropped from three states. Thus, the size of the supplemental sample was reduced to about 11,000 households in 155 PSUs.


January 2003. The 2002 Census Bureau occupational and industrial classification systems, which are derived from the 2000 Standard Occupational Classification (SOC) and the 2002 North American Industry Classification System (NAICS), were introduced into the CPS. The composition of detailed occupational and industrial classifications in the new systems was substantially changed from the previous systems, as was the structure for aggregating them into broad groups. This created breaks in existing data series at all levels of aggregation. Questions on race and ethnicity were modified to comply with new federal standards....

So before Jack Welch and Stuart Varney want to talk about “mistrust” in unemployment numbers and other such unsubstantiated mush, they’d do well to steep themselves in PSU trends over the years.