A+ R A-

The BLS and that 7.8% Unemployment Rate

E-mail Print PDF

Share this article with a friend

By Brad Furnish

On Friday, Oct. 5, 2012, the Bureau of Labor Statistics shocked the world of business, finance and economics by declaring that unemployment in September fell to 7.8% from its August level of 8.1%. Coming just a month before the Nov. 6 Presidential election, this dubious finding gave rise to widespread rumors and even accusations of collusion between the Obama campaign and the Bureau.

It is highly unlikely that the President's men conspired with Bureau officials or employees to invent or alter employment data. The Obama campaign would have too little to gain relative to its potential loss. A 0.3% fall in the unemployment rate hardly delivered a passkey to the Oval Office, but proof of a collusive conspiracy would doom Obama's chances - even if undiscovered until after the election.

That does not mean there was no wrongdoing here - far from it. It should be obvious to all that something was indeed terribly wrong with the way this episode was handled.

Start with the fact that BLS's unemployment number is obviously wrong. "Wrong" does not connote fraud or deceit. It means there was a short-circuit in the chain of reasoning leading to 7.8% unemployment. The BLS household survey estimated a seasonally adjusted total of newly employed workers exceeding 873,000 in September. This was the largest total since June, 1983 (when annualized GDP growth was 9.3%) and January, 1990 (when growth was 4.2%). (Two larger totals in 2000 and 2003 are not comparable because of data adjustments made by BLS.) Yet our current reported growth rate is 1.3%. Moreover, some 582,000 of those estimated workers were part-time.

The BLS also conducts a survey of job growth based on a sampling of business payrolls. This estimate has declined in each of the last three months, from 181,000 to 142,000 to September's total of 114,000. It isn't just that the two surveys differ - they often do. But not by this much. And the fact that recent payrolls detect falling job growth contrasts starkly with the September household survey's stratospheric increase.

Private-sector forecasters each have their own individualized reasons for thinking that the BLS has lost its marbles. My firm, Access Advertising, is a national leader in placing truck-driver recruiting ads. We compile the Driver Recruiting Index (DRI), which estimates the demand for commercial drivers in real time by sampling classified driver ads in 32 major-metropolitan newspapers throughout the country. The DRI has been declining on a year-over-year basis throughout 2012 and by 10-20% since June. In September, the DRI fell by an average of over 14% compared to August. In order to believe the 7.8% figure, I would have to believe that the agency has detected a major economic expansion from which trucking - which carries two-thirds of the nation's freight - is somehow excluded. This is wildly unlikely.

Here is a sampling of characterizations of the 7.8% unemployment rate and the 873,000 job gain in the household survey: "Must be an anomaly;" "statistical anomaly;" "just a fluke;" "statistical quirk;" "implausible;" "almost certainly a statistical fluke;" "huge statistical outlier on the upside;" "not reality;" "an aberration." All of these comments come from respected economists, forecasters and consultants, one of whom is a former director of the Congressional Budget Office. Some of them are known to be supporter of the Obama administration. None are rabid anti-administration partisans.

Suppose you are Hilda Solis, Obama-appointee as head of BLS. It is 8AM, Oct. 5. You have just been handed the report including the 7.8% unemployment and 870,000 household survey figures. You should: 1. order a double-check of all relevant data generation and calculation. 2. (assuming your results checked out) accompany a press release with an announcement that your sampling procedures produced one estimate that defies common sense. 3. Advise the public that no weighty conclusions be drawn from the 7.8% unemployment estimate, since it is highly suspect. 4. Invite scrutiny of your methods, results and checks by any interested parties.

What did the BLS do? Apparently, none of the above. Instead, they simply released the results with no special emphasis. (Later, Ms. Solis even refused to cast doubt on the suspect estimates and even defended them, somewhat obliquely.) This had the effect of inviting the general public, the press and even sophisticated analysts to take the information at face value. Quite a few people did. Others viewed it as a slap in the face. And some of those started to question the honesty of the whole process.

Why did the BLS do nothing instead of the right things? Their inaction may have been due to political bias or to bureaucratic inertia. In this case, the outcome of a Presidential election might ride on the procedures for handling a foul-up like this.

But the point would apply just as strongly if the data had been released on Nov. 7 or any other day. There are always important matters riding on the content of government data releases, and people want to believe that government employees are acting in good faith and providing reliable information. The best way to convince people that the government is here to help them is to go out of your way to do just that.

You are not currently authorized to post comments.

Quantcast