April 14, 2006

Seeking to make algorithmic trading even more predictive and less reactive, Wall Street brokerage houses and quant shops are examining real-time news as a feed for their trading models. Algorithms always have been based on the numbers -- such as price movements, volumes and volatility. But now, companies have started to look for other sources of information they can model, according to industry observers.

Over the past year, there has been a lot of buzz in the industry around the idea of filtering news feeds based on events such as bankruptcies, executive changes, and mergers and acquisitions, as well as other so-called event triggers that can impact future stock performance. Clearly, breaking news can give traders an edge -- if they act on it faster than the competition.

Quantitative trading groups say that in the near future they will look to develop models that use the historical impact of news events on stock performance to predict the effect real-time events may have on future performance. And, potentially, Wall Street quants could develop programs that automatically trade off the real-time news feeds. In fact, some experts predict that the next generation of algorithms could be based on news events. They've even coined the term "news-flow algorithms." >>

"It will be another factor that we would incorporate into the models," confirms Peter Cherasia, CIO and senior managing director at Bear Stearns, where he also is cohead of the Equities and Analytics Systematic Trading (E.A.S.T.) division.

To get in front of the trend and provide traders with the data they need to develop such algorithms, business content providers, including Dow Jones & Co., are releasing historical news archives and real-time news feeds. Reuters also is believed to be working on news products for algorithmic trading, though it has not released specifics.

"News is the new frontier," says Kirsti Suutari, global business manager for Reuters' enterprise division. "The people who are very rehearsed in algorithms and strategies are looking for the next competitive edge."

But Suutari notes that news-based algorithms are not a done deal. "You may find that among the market leaders there will be some divided opinions," she offers, suggesting that news may provide additional input for algorithmic trading rather than lead to a new kind of algorithm. "Maybe news will be useful as a risk management feature," she says. "The algorithm can be chugging along nicely, then a news item comes along and skews the market. As soon as the market starts to behave differently, the algorithm needs to know when it's no longer good to continue the trading pattern."

Already, sell-side firms have begun subscribing to services that filter news feeds and other documents -- such as SEC or regulatory filings, corporate Web sites and blogs -- as another data source they can plug into their quantitative models. "It's only in the recent past that companies have started to look for ways to anticipate what's happening in the marketplace," says Len Dreifus, director of technology at Dow Jones Newswires. "They're looking for news as one of those components."