Google blows up UAL

Let's avoid the giant elephants in the room for a moment (Fannie Freddie Lehman Merrill), so I can point out another technical folly that has come out of the woodworks.

On Monday, UAL's stock price was essentially wiped out this week by an error in the systems that track news stories on the internet.

Here's a breakdown I'm citing form the LA Times:

Back in 2002, United was verging on bankruptcy, and South-Florida Sun Sentinel newspaper, a subsidiary of Tribune group, published an article about it. The Google News search algorithms came across the article early Sunday, and by opening it, probably pushed it to the "most viewed" section of the webpage. From there, other news aggregators that just search over "most viewed" lists picked up the article and, due to a mis-dating on the original article, presented it as current news to Income Securities Adivsors, inc. Bloomberg news picked up the "story" from there, and automatic trading algorithms that are programmed to respond immediately to certain news events, like bankruptcies, dumped shares, cutting away 75% of UAL's market cap before trading was manually frozen.

Things are back to normal now, but this brings to light some pretty interesting considerations, as the effects that automated information system processes have on our lives increase.

These systems work fine, as long as they are designed perfectly, but when you incorporate human error, like mis-dating an old news article from a tiny newspaper, the shock waves that can be made are staggering. This is especially true in capital markets, due to the affect information has on liquidity. In spite of all of it's value in maintaining fair value, liquidity also has the adverse effect of draining a security of value, whether the information causing the change is accurate or not. 

How will we solve these issues? Insurance first comes to my mind as a "solution," but it exists more to treat the symptoms, not the cause. I imagine that more and more adaptable circuit breakers will need to be put in place. We will need to find innovative new solutions that fight to maintain a fair market value that still has a solid, auditable foundation of value. Perhaps we can use historic data to statistically model past market responses to news items like bankruptcies, and design a circuit breaker that prevents a stock's price from swinging too far past that average?

Maybe such intervention sounds too manipulative, even for these times. Speaking of market manipulation, I'll try to throw something together about the big news stories soon.