At 9 a.m. ET Friday, Labor Secretary Hilda Solis announced that the Bureau of Labor Statistics' survey found only 18,000 new jobs were created during June and that the nation's official unemployment rate had ticked up a tenth of a percent to 9.2%.
The Dow, Nasdaq and S&P indexes all dropped like rocks at the 9:30 a.m. opening bell, falling by more than 1% each. By the end of the day, they had recovered some ground, but the Dow and Nasdaq were still down by almost half a percent and the S&P shed 0.7%.
Most analysts and news reports attributed the initial plunge in equities markets to the discouraging jobs number which was well below the CNN consensus forecast (120,000 new jobs) by business economists. Other consensus figures ahead of the BLS release were in the 100,000 range, with Moody’s Analytics expecting 100,000 new jobs and S&P’s economists expecting a 90,000-job bump.
What may be surprising to some, however, is that all the consensus forecasts could actually turn out to be right because when the final, revised numbers are finally tabulated (almost a year later), all three estimates fall within the report's historical margin of error.
But that didn't stop investors from selling off shares in early trading Friday.
“It is kind of funny when there is such a dramatic reaction in the market,” Betsey Stevenson, chief economist at the Department of Labor, said in an interview with On Wall Street.
She explained that over the past 35 years, the monthly jobs figures have been routinely revised and adjusted after the initial release and end up being, “on average,” around 115,000 jobs higher or lower than originally reported.
Ryan Sweet, a senior economist with Moody’s Analytics, conceded that being off in his forecast of the new-jobs estimate by 72,000 is “a humbling experience,” but he agreed that it is not actually a significant difference. “Basically the BLS has a confidence factor of 100,000 jobs,” he said, “so the actual number of jobs created could be 100,000 higher or lower than what they initially report.”
Why the dramatic investor reaction then?
"It’s often overlooked by investors and analysts that the number can be off by that much either way," Sweet said.
But what about the ADP jobs report for June? It was released -- as usual -- a day ahead of the BLS number and it reported 157,000 new jobs for the month.
Joel Prakken, an economist with Macro Advisors, which runs the numbers for ADP, said that his survey doesn’t include public employment -- which the BLS does -- and that sector, according to the BLS, lost 39,000 jobs last month. Take the public jobs out of the BLS survey and you get 57,000 new private sector jobs, putting the ADP number also within the BLS (and the ADP) margin of error.
Both the BLS and the ADP, by the way, do “adjustments” to their hard survey numbers.
The BLS does a seasonal adjustment to account for the usual increase in new jobs created in June compared to April or May due to new and often temporary summer jobs. Stevenson said this adjustment decreased the actual total number of jobs in the U.S. (132,079,000) by 1,062,000 jobs this time around.
ADP, for its part, adjusts its jobs number to account for the fact that its 400,000-plus company survey includes only self-selected firms that use ADP services. The BLS survey group, also of more than 400,000 companies, is randomly selected.
"While the markets hang on every 1,000 jobs in each BLS report, and I understand that it’s always the first piece of economic information out each month, I always end up trying to remind people that coming up with a jobs number is a difficult sampling process with wide margin of error," Prakken said.
The DOL’s Stevenson agreed and said she too was surprised by today’s meager 18,000-job increase. The BLS staff runs a tight ship, keeping the jobs and unemployment numbers secret, even from Solis, until an hour before the official announcement.
“I thought it would be 100,000 myself," Stevenson said. "But one monthly figure doesn’t tell you much. These jobs numbers provide a pretty good measure of what’s happening in the economy, but you really need to look at trends over 12 months’ time.”