While this will hardly come as a surprise to any of our regular readers, occasional visitors may be confused to learn that according to a discovery by the Carolina Journal, North Carolina "Gov. Bev Perdue’s press office has received access to confidential employment data from the U.S. Bureau of Labor Statistics hours if not days before its scheduled release, quite likely in violation of federal law." Once again the rabbit hole, which these days is pretty much everywhere, emerges: "Documents and correspondence obtained by Carolina Journal show that the Division of Employment Security, formerly known as the Employment Security Commission, sent a draft of the press release each month to Perdue’s press office. The governor’s spokesmen typically rewrote the text and added a positive spin, even if the data did not support Perdue’s talking points." And while one may say this is a perfectly innocuous leak of otherwise embargoed data, others may highlight the following facts: "While the operation may sound like a harmless effort to add political spin to the release of jobs data, sharing confidential BLS estimates while they are protected by an embargo violates a federal law barring the early release of employment data. This is no small matter: A conviction for breaching the Confidential Information Protection and Statistical Efficiency Act of 2002 carries a fine of up to $250,000, up to five years in prison, or both." Of course, when it comes to breaking the law, both members of the US banking class, as well as America's politicians, are perfectly immune from any repercussions. But at least the next time the market does its usual pre-NFP acrobats, the only question will be: which particular US politicians i) traded in advance of the embargo lift, and ii) leaked the information to ten of their closest friends, who did the same, who did the same, etc.
Furthermore as it turns out, this is not the first time the North Carolina governorship has been caught red handed.
Former Gov. Mike Easley’s administration received an early briefing on the employment reports from ESC staff for several months in 2003 and 2004, if not longer, but those briefings apparently stopped after CJ learned of and reported on them in 2004.
But why stop if nobody was at fault, unless....
As to the particulars of this particular case:
Documents and correspondence obtained by Carolina Journal show that the Division of Employment Security, formerly known as the Employment Security Commission, sent a draft of the press release each month to Perdue’s press office. The governor’s spokesmen typically rewrote the text and added a positive spin, even if the data did not support Perdue’s talking points.
The glowing quotes were attributed to Lynn Holmes, director of the employment agency, but the documents show the quotes were approved and probably written by a Perdue press aide, either Chrissy Pearson or Mark Johnson.
In several instances, DES spokesman Larry Parker cautioned Pearson or Johnson against using extraneous or unverifiable information in a release to boost Perdue’s image. At times, the Perdue communications team would push back, and the release would undergo several revisions before final publication.
Perdue and her staff may not receive formal briefings, but Parker told CJ in September that he shares the jobs report with Perdue’s staff roughly 24 hours before the announced time for the monthly release. Emails obtained by CJ involving Parker, Pearson, Johnson, and in one instance Department of Commerce spokesman Tim Crowley, show officials discussing the data more than 48 hours before the embargo was lifted.
Enter the BLS with the form denial of everything:
“BLS does not support the release of employment and unemployment data before the established dates and times,” said Janet Rankin, BLS regional commissioner for the southeast office in Atlanta in an email. She notes that states are required to publish the dates they will release employment data on Dec. 31 of the previous year and follow that schedule.
She added that state agencies enter a cooperative agreement with BLS to gain access to its confidential data, and that the agreements “state that estimates cannot be shared outside of the LMI unit until they are final and ready for publication. The data are embargoed until the state predetermined release date and time” (her emphasis).
And the smoking gun:
When CJ reported on the briefings Easley received in 2004, Rankin said at the time: “A person associated with developing the data that is caught releasing it or commenting on it prematurely is subject to fines and jail time.”
That's the BLS essentially saying that one of its own employees is committing a felony offense. Over and over. And here we are worried about statistical impossibility of the endless prior upward revisions in Initial Claims, when the fraud runs to the very core of the reporting system. Why for the hardened BLS criminal, the least thing they can and will do is fudge data mercilessly and without fear of consequences. After all, there are "birth/death" adjustments.
The rest of the article, which is worth a read in its entirety goes into details of just how the CJ caught the current governor. And for those forensically inclined, the trail of emails can be found here.
And of course, by simple logical induction, what happens in North Carolina, happens everywhere else. So good luck to all the folks trying to trade in "advance", or just after the release of the 8:30 am NFP number.
What is most ironic in this case, and likely in every other one which involves collocated High Frequency Trading boxes to "front run" the rest of the slower market upon critical BLS data (which is virtually half of the Class A high frequency economic data): all that money for high tech vacuum tubes was spent for nothing. At the end of the day, the slowest trader, who has the best connections with the BLS, ends up trumping the guy with the $5 million a month box parked squarely at the receiving end of NYSE's Mahwah fat pipe.