You Really Don't Want To Hitch Your Credibility To Jack Welch

Yesterday's drop in the U-3 unemployment rate to 7.8% was shocking, given the absence of a commensurate improvement in the underlying economy, or indeed in the broader labor market as measured by the U-6 rate, which was flat at 14.7%.

Coming on the heels of Obama's weak debate performance Wednesday night, this seems to have led a lot of people to speculate about whether the Obama administration has somehow compromised the integrity of the BLS for the benefit of their re-election effort.

While I wouldn't rule such a possibility entirely out of the realm of possibility - I am on record as stating that Barack Obama is not merely a bad President but a bad person - I think it is really, really unlikely.

And the grandstanding of Jack Welch makes me even more skeptical than I would otherwise be, about which more in a bit.

One good rule in the blogosphere is: when in doubt, trust Megan McArdle. And she delivers a good overview of why the most likely explanation for the U-3 outlier is entirely innocent: the inherent volatility of the household survey.

Conservatives dismayed by the misleadingly bullish headline unemployment number should focus less on the probability that the numbers were cooked, which is quite low, and more on the probability of mean reversion in the November number, which is high. If the September U-3 number is truly the outlier it appears to be, then there is every reason to anticipate that the October number, which will be released four days before election day, will rebound to 8% or higher.

I do think one argument that Megan (and others) have advanced is faulty, and that is the idea that there is no rational motive to manipulate the data because it is too late in the campaign to do Obama any good. People making this argument invariably cite the experience of George H.W. Bush, who lost despite improving unemployment numbers as the 1992 election approached. I think this overlooks the Feiler Faster thesis popularized by Mickey Kaus. Voters will receive and process information much faster in the 2012 campaign than they did twenty years ago, and so Obama could very well benefit from late trends that did not help George H.W. Bush.

The worst thing about this episode, though, and the one with the potential to do the most long term harm to the conservative cause, is that it has enticed people into ascribing credibility to former General Electric CEO Jack Welch, who has been publicly questioning the integrity of the unemployment data.

When the issue is credibility, history suggests that you should be very wary of lining up on the side of Jack Welch.

A little over 10 years ago I started looking into General Electric. The consistency of the company's earnings over many years under Welch had been an anomaly. The idea that so large a company could grow earnings so consistently amid the vicissitudes of the global economy seemed like a red flag.

In reality, this was evidence of what is described in boardrooms as "earnings management," and outside of boardrooms as "fraud." Earnings management doesn't always rise to the level of fraud in the legal sense (though it can), but in the colloquial sense that's all it is; the practice is intended to hide the true performance and condition of a company, including from the investors who own it.

When asked about the seeming improbability of GE's earnings trajectory, I found that executives often fell back on a mantra: "We manage companies, not earnings."

Unfortunately for them, it was about that time that Jack Welch published his quasi-memoir, "Jack: Straight from the Gut." On page 225, Welch described the aftermath of the Joseph Jett trading debacle at GE-owned Kidder Peabody, in which it was suddenly discovered that Jett had lost the firm $350 million:

"That Sunday evening, I called 14 of GE’s business leaders to deliver the bad news and apologize to each of them for what had happened. I felt terrible, because this surprise would hit the stock and hurt every GE employee. I blame myself for the disaster.
 
The previous year, 1993, when Jett’s phantom trades accounted for nearly a quarter of the profits made by Kidder’s fixed income group, Jett had been named Kidder’s “Man of the Year.” We had approved Mike’s request to give Jett a $9 million cash bonus, a huge award even for Kidder. Normally, I would have been all over this. I would have dug into how one person could have been so successful, and I would have insisted on meeting him.

The response of our business leaders to the crisis was typical of the GE culture. Even though the books had closed on the quarter, many immediately offered to pitch in to cover the Kidder gap. Some said the could find an extra $10 million, $20 million, and even $30 million from their businesses to offset the surprise."

That is the textbook definition of earnings management, if not outright fraud, in Welch's own words. Barry Ritholtz offers a good roundup of other information that reflects upon Welch's leadership and credibility here.

In the battle of ideas, your credibility is really all that you've got. It's the reason why I made such a big stink about Reuters' faulty claim that "GM is still losing as much as $49,000 on each Volt it builds." Many smart people, who believe as I do that the auto bailout and ongoing Volt subsidies make the nation worse off, impaired their own credibility and that of our shared cause by embracing Reuters' faulty argument.

Writing about another issue entirely earlier this year, Megan McArdle offered a singularly valuable bit of wisdom to her readers:

"After you have convinced people that you fervently believe your cause to be more important than telling the truth, you've lost the power to convince them of anything else."

That's true for Barack Obama, and its true for Jack Welch. 

Anyone wondering how much credibility to ascribe to Welch's comments on the unemployment data should bear that in mind.