I try to avoid quoting other people’s opinions and articles on this blog, but over the weekend I found a newspaper article about the interaction of search engines and the stock market that was absolutely fascinating. It also seems like an appropriate topic with this week’s events on Wall Street.
The story is that trading companies have built large search engines that scour news reports and actually use the results to make automatic trades. The article starts by talking about the use of computers in the market:
“Automatic stock trading based on algorithms that react to market moves and conditions have been a feature on Wall Street for years ? nearly 50 percent of all trades on the New York Stock Exchange are expected to be automated by 2010, according to the Aite Group, a Boston-based consulting firm, up from roughly 30 percent in 2006.“
I have a friend who was a successful day-trader for years until 2007. He told me the story of how he walked into a trading company in Charleston, South Carolina. The person showing him around pointed to a large server farm and said “those are our day traders”. The servers were executing 1000s of trades every second. He said it felt like a telegraph operator seeing a telephone for the first time and quit day-trading the next day.
A News Search Engine for Trading
The article goes onto explain the origins of trading based on news and search engines:?
“Robotrading based on news events is of more recent vintage. ?It?s growing exponentially, but nobody wants to talk about it because they feel they have an edge,? said Professor Lo, who directs the M.I.T. Laboratory for Financial Engineering. The formulas are viewed by banks and hedge funds as trade secrets, he said.
“Many of the news-based trading formulas today have as their genesis research Professor Lo conducted at M.I.T about five years ago when he sought to measure the ?emotional context? of stories in The Wall Street Journal. Professor Lo chose words such as ?anxiety? and ?depression? and counted how many times they showed up in the newspaper on a particular day ? and then looked for correlations to the stock market.
?How many times they appeared on a given today tended to be a leading indicator of the S. &P.,? Professor Lo said. From there, he developed more complex methods such as the ordering of words ? pairs and triplicates. ?We were among the first to use algorithms to read the news,? he said. The formulas assign a weighted value to certain words ? ?bankrupt? may be given a higher score than, say, ?anxiety? or ?down? ? and when a certain level is reached, depending on how the model is programmed, it will automatically initiate a trade .”
The newspapers are happy so the market goes up. The writers are pessimistic so the market goes down. It goes even further:
“Even the news organizations themselves are in on it, developing support tools so traders can more quickly mine their news content to make trades. Professor Lo has been collaborating with Thomson Reuters to develop news-based algorithms. Dow Jones and Bloomberg offer similar services. “
So the news organizations are not only able to actively moving the markets with their opinions but they’re also providing the tools to encourage it. Sounds like a recipe for insider trading by journalists … what would stop a journalist writing an article about a company, including the word “bankruptcy” ten times and then immediately short selling the company’s stock?
“The important factor is speed ? receiving the data, analyzing the information and making a trading decision in a matter of milliseconds.“
What could possiby go wrong with such hasty decisions? Ask United Airlines who had $1 billion wiped off their stock value in 12 minutes because of a story from 2002 falsely entered in Google News.
“It doesn?t stop with hard news. Services exist that track sentiment for a company on blogs and social networks on the Web.“
So not only do newspaper experts have the power to shift the markets, but so do bloggers.
Some Takeaways
- The idea that opinion creates reality isn’t new to those interested in politics, movies , or even physics. As we gain more outlets to express our opinion, people will be able to measure and act on them.
- Don’t be surprised to see search engines reach beyond the internet and play a larger and larger part in everyday life.
- Search engines may increasingly look for patterns rather than specific results. After all, what is the U.S. government’s electronic surveillance system but a large search engine that is sometimes used for precise information?