After years of sitting on the sidelines, the buy side finally is in the algorithmic trading game. Algorithmic trading, or computer-directed trading, not only cuts down transaction costs, it also allows investment managers to take control of their own trading processes.
Although hedge funds, many of which use highly quantitative trading strategies, have used algorithms for years, traditional buy-side firms now are being targeted as the best opportunity for algorithmic trading growth. In fact, Aite Group, a Boston-based consultancy, expects traditional buy-side firms to account for 30 percent of all algorithmic trading by 2008 -nearly double the current figure.
With that growth comes opportunity. The buy side currently is being bombarded with solutions for algorithmic trading implementations. Everyone from bulge-bracket firms to agency brokers to vendors is offering some alleged remedy, creating confusion for a marketplace full of asset managers unseasoned in the trading strategy.
Uncovering the 'Secret Sauce'
"There are so many different [algorithmic trading] options out there right now," observes Sang Lee, founder and managing partner at Aite Group. And while a benchmark to measure those options against each other would be ideal, Lee says, "For that to happen, all the firms that are providing them would have to agree to give you their algorithms to test them. But that's the secret sauce - it's their competitive edge. They will never do that."
As a result, many buy-side firms are doing the only thing they can when the answers are not readily available in the marketplace: building algorithms themselves. Ohio Public Employees Retirement System (OPERS), a Columbus, Ohio-based firm with $64 billion in assets ($26 billion in equity), began using algorithmic trading about 18 months ago for its Russell 3000 index funds. "Cost control is paramount to your success in index funds," says Joan Stack, trading manager for OPERS. "Employing rules-based strategies has enabled us to increase productivity, lower commission costs and reduce our implementation shortfall."
Stack chose to implement a front-end broker-neutral platform from Great Neck, N.Y.-based FlexTrade that integrated with the firm's current order management system from Boston-based Macgregor. The off-the-shelf FlexTrade system came with "canned" algorithms, explains Stack, such as VWAP, transition trading, pairs and long/short trading. Gradually, though, Stack's team began tweaking those algorithms to accommodate for OPERS' particular strategies, and finally the firm hired an academic to work with the traders to write proprietary algorithms. Right now, she notes, about 5 percent of the firm's algorithms are proprietary.
Stack understands that OPERS is lucky to have a budget that enables the hiring of talent to write algorithms. "Not all buy-side shops have the luxury of being able to write their own," she says, conceding that it's necessary to have an internal IT department capable of spending time with traders to understand their strategies, then coding algorithms. "A lot of the off-the-shelf systems are good, as long as you understand what they're trying to achieve."
Barclays Global Investors, based in San Francisco, also employs an internal team to create a small number of proprietary algorithms. Richard Tsai, BGI's head of electronic trading, explains, "The main advantage is having clear understanding over the entire investment process. Our traders know the life cycles of their trades and have a clear sense of how that order should be executed."
BGI uses an internal team of about 20 people, called Trading Research, to establish the appropriate trading strategy, assist in trading and interpret transaction quality. "We're always looking at ways to improve the execution of orders," Tsai says, and that often can be through an internal algorithm. "We feel we have certain insights, based on the level of the activity that we do in the market, that we would like to retain."
However, Tsai points out that more often than not, the firm turns to its sell-side partners to provide algorithmic trading, as well as other electronic trading strategies. Tsai determines which partner to turn to based on best execution. "We do witness from time to time that certain strategies are not exactly what was advertised," he says. "If you do your due diligence correctly, you will witness differences" in partners.
Innovation is a discerning factor, too, he asserts. "As strategies develop, there is a constant evolution and escalation process. Somebody always can come up with a better mousetrap," Tsai says.
OPERS' Stack agrees that innovation makes algorithmic partners more attractive. For example, she says, an algorithm that could balance an entire portfolio instead of single stocks would be helpful, as would one that could handle other asset classes, such as real estate investment trusts.
Pre-trade analytics also are a selling point, Stack adds. They often enable the buy-side trader to become more comfortable with algorithmic trading by providing expected results of the trade.
Harrell Smith, an analyst with Celent Communications, believes pre-trade analytics, as well as other types of education from the sell side, will help buy-side traders move toward greater adoption. "The buy side is not educated as to the potential uses of these algorithms," he says. "You could pick up the phone and spend five cents per share, or do it yourself on a screen and pay a penny per share. But if you don't know what you're doing, you're not doing yourself any favors by saving that money."
Despite its perceived universality, right now, algorithmic trading on the buy side is more about hype than actual demand, the Aite Group's Lee contends (see sidebar, at right). And due to a lack of education in the marketplace, many buy-side firms are acting on fear, he adds. "They see their competitors talking about algorithmic trading and think they need to get in on it," Lee says. "The adoption rate is not as aggressive as people think it is at this point. It's all about execution. Algorithms are just one option."