Be Sure You Trust Your News Source For Old News Dressed As New Could Cost You Big-Time

Philip Stone September 10, 2008

How many times have you searched Google looking for recent information but clicked on “web” instead of “news” and only by chance did you notice the item was a few years old. Now what if that old item actually was found by Google on a newspaper site and the robot, not finding a date on it, listed it as new news item? Scary? Couldn’t happen? Well it did last weekend and it caused financial mayhem in United Airlines shares Monday.

The Google spider crawling the South Florida Sun-Sentinel’s web site found a six-year old Chicago Tribune United Airlines story announcing it was going into Chapter 11 bankruptcy. The story was said to have been in the archive, but it did not contain a publication date so the robot apparently assumed it was a new story and gave it that day’s date within Google search. What then seems to have happened is that the story got “Googled” enough times that it made it onto then “Most Read” section of the site since and with Google’s logarithms programmed as they are the more often a story is accessed the higher it’s priority in the search engine.

Come Monday and a reporter for the Income Securities Advisors, a Florida investment newsletter, does a Google search on “bankruptcy 2008” and the first item that pops up is that old Sun-Sentinel story now dressed as new. So the reporter “assumes” the story is new (a casual reading of the article should have sent up warnings the information was old but it got through as new)  and since the newsletter posts its information on Bloomberg Terminals that’s where the story appeared for the financial world at large. It’s probably best to rely on Bloomberg’s own story for what happened next. “Income Securities distributed the report on the Bloomberg Terminal before retracting it and issuing a correction.  Bloomberg News also ran a headline citing the Tribune story after the article appeared on the Sun-Sentinel Web site.”

Regretfully, we don’t have access to a Bloomberg Terminal so it’s difficult to sift through the muddle of actually what happened at Bloomberg. Its own version of the story says that Bloomberg News ran a story separate from the posting by the Florida newsletter, yet a Bloomberg spokesman told Reuters that since the newsletter story was not written by Bloomberg, it was not edited by Bloomberg. What may have happened is that while the full newsletter story goes only to those subscribers who pay to receive it via their Bloomberg Terminal, non subscribers will also see a two-line summary – so is that the news item referred in Bloomberg’s own story or was there an additional separate Bloomberg story? 

Richard Lehmann, president of the newsletter company, is adamant that the Sun-Sentinel story got picked up because it had no date on it and it appeared next to current news about Hurricane Ike.

As for Google, a spokesman points out, "The article first appeared on the Sun-Sentinel site, independent of Google News.”  On its blog site Google added that as soon as it was notified the link was in error the story was immediately removed. "It has been widely reported that many readers were unable to determine the original date of publication of this article, and our (automated computer) crawling was similarly unable to recognize that the article was old," Google stated.

So there we are: no one did anything wrong; it’s just an unfortunate set of circumstances. Go tell that to the investors who lost big-time. And you wonder why there are so many lawyers in the world!

Anyway all it took was the headline “United Airlines files for Ch. 11 to cut costs"  to cause, within seconds, financial mayhem in not only United shares – about $1 billion got wiped out  as shares plunged from $12.30 to $3 -- but also other airline companies got nailed, too. United seemed to take a while to figure out what happened but it finally released its  “not true” statement and by the end of the trading day the shares recovered to $10.92 – still off 11.22% on the day. On Tuesday the shares fell another 4.67% to close at $10.41, but that probably had more to do with rising oil prices than news from six years ago.

Automation may also have been responsible for much of Monday’s sell-off. Many investors have “stop-loss” orders on their holdings – if the shares drop below a certain price they get automatically sold. Investors can complain later if systems went wrong and the stock exchanges have the power to bust such trades, but in this case the exchanges say the trading was orderly and they are letting things stand.

But to give an idea how financial news can spark volatility, around 15 million shares traded during the 13 minutes until trading was halted (which again goes to show why 15 – 20 minute delayed quotes available on the Internet are useless and why recent  real-time upgrades by Yahoo and Google to their financial services was so crucial).

This all has to be “hog heaven” for the lawyers. Who’s to blame? NASDAQ, the main trading exchange for United shares, says all trades count even if they were made based on wrong information. Thus those investors who dumped United shares at, say, $3, when they started the day above $12 are just plain out of luck, unless, that is, the lawyers can nail somebody.

David Herron, president of the Chicago Stock Exchange summed up the stock exchange position. “We don't police the information that's out there. People take risks when they trade on news." In other words, one should be very sure of one’s source.

So is the newspaper liable that a six-year old story was able to be picked up by the Google robot without identification that it was an archive story? Could the newspaper argue, “Who invited Google, anyway?” Is Google at legal risk for having listed the story as current news?  Is there liability at Bloomberg for third parties putting out wrong information on its system (want to take a bet on how tight their liability language is on things like that?), and did Bloomberg carry its own news item based on the Sun-Sentinel story?

What is of particular interest here is how much of this was all down to automation and how much was human error. The Chicago Tribune quotes Joe Schwerdt, deputy managing editor-interactive for the Sun-Sentinel, saying internal tracking records prove no one at the paper had opened the original story file since 2003. The story would have been available via a site search, but no one outside the paper should have had access to the story file, Schwerdt said.

SmartBrief.com says the story was posted to the Sun-Sentinel’s site at around 1 p.m. Sunday. If it was an archive story how could that be unless reaching “Most Read” status brought it automatically to the forefront? It took about 23 hours to pull the story and that got done only after Schwerdt  said he was alerted after a Tribune phone call to Sun-Sentinel editor Earl Maucker (would have been fun to have been the fly on the wall listening to that one!).

"I literally just got word a couple minutes ago that there was problem," Schwerdt said when first asked Monday.  He said he had no idea how the old story could have been on the site looking new.  And Tribune lawyers, no doubt smelling the upcoming litigation, are adamant that the story had not been reposted as if new, and that it was located in the site’s archive section.

“The story contains information that would clearly lead a reader to the conclusion that it was related to events in 2002,” a Tribune statement said. “In addition, the comments posted along with the story are dated 2002.” Well, obviously all of that wasn’t so clear, for whatever reason, to the newsletter reporter.  

We know the Google robot is automatic so the questions there are what is it able technically to pick up, and is the Sun-Sentinel archive behind any firewall that prevents Google access?

And what happened at Bloomberg? Sure, its editors keep hands-off a posting by a third party. But did Bloomberg have its own news item on the bankruptcy as its own story says, or is it a matter of semantics and there was just the third party item? If it did have its own item, as its own story claims, then why did a Bloomberg editor fail to ask out loud, “If this happened over the weekend and this is Monday then how come no one else has it”– red flags should have gone up – or was automation somehow involved?

It all brings back the old BBC World Service policy of never using a non-BBC sourced story unless it had at least two sources – for international affairs that meant two news agency reports. This is as good an example as any of why that’s not a bad rule to still follow today (for those who still have the luxury of subscribing to more than one news agency!)

But in the financial world where a second or two can make the difference in making or losing a fortune the money people won’t wait around for confirmation. Which means you’ve really got to trust your source. And in this case it was information by a third party, but it was delivered on the Bloomberg Terminal.  Guilty by association?

Competition between the financial news agencies to be first with news is fierce.  Workloads and pressures are great, mistakes can happen, but as this fiasco showed financial news mistakes can costs a billion of whatever currency you would care to name. The old news maxim that you need not only to be first but also right still holds sway.

Bloomberg seems particularly to be having a hard time of it these days. Last month it inadvertently ran a Steve Jobs obituary but luckily nothing happened with Apple shares as the US markets had closed 30 minutes before. It also ran an incorrect report that Sarah Palin had been arrested for drunken driving 22 years ago because its reporter misread a New York Times story; it was, in fact, her husband who got stopped. Bloomberg quietly made the correction.

But in this case, where around $1 billion of capitalization got wiped out within 13 minutes based on wrong information, it’s not so easy to sweep it all under the carpet. There’s a reason why news agency liability clauses are carved in stone, especially for delivering third party news, but this time the lawyers are going to be kept busy for a long time to try and pinpoint financial blame.