Does the Margin of Error Make the Jobs Report Meaningless?

There's a margin of error of +/- 100,000 on that payrolls figure. Here's why the report is still useful.

By SHARE
140207_jobs1

Everyone seems to agree the latest monthly jobs report was a disappointment. The 113,000 jobs that employers added in January was way below expectations and seems to confirm that the U.S. has had two straight months of pretty ugly job growth.

Except wait. According to the Labor Department, there’s a margin of error of 94,600 jobs on the latest establishment survey, which produces the monthly nonfarm payrolls figure. In other words, the Bureau of Labor Statistics is telling us there were somewhere between roughly 18,000 to 208,000 jobs created last month.

[READ: How Twitter's User Growth Compares to Facebook's]


That’s a huge range, and it seems to imply that all of the handwringing over any given jobs report is overdone. Had the Labor Department reported 208,000 jobs, it would have been heralded as a remarkably strong month for job growth. And had the survey registered a number of 18,000, it would have been reported as disastrous.

Except don’t dismiss it yet. The Labor Department is saying there is a fair amount of uncertainty in that estimate, one department economist says, but he adds that 113,000 is still the Labor Department’s best guess, and with good reason.

“The farther away you get from 113 [thousand], the less likely it is to be,” says John Coughlan, a staff economist at the Bureau of Labor Statistics. “So it's pretty likely it could be between 100 and 123, or something like that.“

It’s not just the margin of error that means that 113,000 figure could be wrong. The monthly establishment numbers go through two more revisions as the Labor Department receives more data from firms, and then the numbers go through another annual benchmark revision after that.

“This is really our first shot at it,” Coughlan says.

So while the Labor Department is giving us its best guess, that guess often can shift over time. Which once again raises the question of why jobs day is such a big day each month.

“You have to take it sort of like a ladder,” says Patrick O’Keefe, director of economic research at accounting firm CohnReznick and a former deputy assistant secretary at the Labor Department. “The most compelling rung on the ladder, being the first one, is each month it's the only data we have.”

[ALSO: Friday Roundup: Jobs Edition]


Beyond that, he says, is the question of how to interpret and explain that number and its limitations. He says it can help to think of monthly jobs estimates not as fixed points but as “fuzzy smudges on a chart.” One smudge indicating a low job creation number might be troubling, but that smudge also might contain a different true measure of jobs created than the central figure the department reported. That smudge also could simply be an anomaly. However, a string of smudges around low numbers would be more troubling.

Given that this was the second straight disappointing jobs report, is it now the time to worry about smudges? Maybe not yet, another economist says, pointing to yet one more way the jobs report can be inaccurate.

“I’d wait until March,” says Michael Dolega, senior economist at TD Economics. While the government’s estimates are seasonally adjusted, they do not adjust for unusual events, like natural disasters and unusual weather patterns that can affect normal business patterns. “Both December and January were extremely cold, and there are factors that BLS doesn't adjust for.”

At the very least, the range of likely job counts around January’s 113,000 estimate does not include zero. That means we can, with a large degree of confidence, say that some number of jobs were created in January – which is good news, just very uncertain good news.