U.S. Algorithm Spending Slows
After growing at substantial rates over the last several years, spending on algorithmic applications by the buy side is slowing in the United States. According to research by Datamonitor, buy-side spending in the U.S. on algorithmic trading applications will rise slightly from $442 million in 2007 to $454 million in 2008 and $469 million in 2009 (see chart, page 30).
While large investment management firms and hedge funds will continue to spend on algorithms, experts predict their spending will slow. And many buy-side traders at smaller shops remain skeptical about depending on algorithms to trade large amounts of their orders, particularly if they are focused on smaller-cap names.
But as we move into 2008, many algorithm providers are hoping to spur renewed interest with new and different offerings. The slowed investment here in the U.S. leaves the door open to increasing growth overseas and a new crop of algorithms that might just boost spending on automated trading strategies beyond its current plateau. Next up in the algo space -- better dark pool access, more-adaptive strategies, portfolio-based strategies and algorithms that go beyond equities into new asset classes.
Though Michael Crockett, senior equity trader at Dallas-based Brazos Capital Management, trades mainly smaller-cap stocks for which he doesn't generally use algorithms, he says the older generation of algorithms is past its prime. Emerging algorithms, Crockett suggests, will be more sophisticated and adaptive. "If someone is still using VWAP or participation algorithms, there is no way they are getting best execution," he contends.
Reflecting an attitude that likely has contributed to the algo-spending slowdown, however, Crockett notes that he is hesitant to use algorithms to a large degree because he depends on his sell-side relationship. "When I put an order in an algorithm, I don't know who the other side is; I don't know how much information I am giving away or how much that could cost me," he explains. "I depend on solid, ethical relationships I can trade on."
Still, Crockett adds that he does "look at every algorithm and every technology I can get my hands on." Unfortunately, he says, there are only a few firms that understand where the market is going and have enhanced their algorithms to reflect the new market dynamics.
Hit the Pools
Some of the biggest trends shaping the markets have been the proliferation of dark liquidity pools and increased trading outside of the U.S. and in new asset classes. As such, experts agree that new algorithms that address these trends will account for much of the growth in the algo space in 2008 and beyond.
Amit Shah, financial services technology analyst at Datamonitor, says the next generation of algorithms will be more focused on advanced stealth trading. "Algorithms will be more advanced in the area of dark liquidity and taking into account any pools of liquidity out there," he asserts.
Joseph Wald, CEO of EdgeTrade, agrees. "What we'll see going forward into 2008 is the adoption of liquidity-seeking algorithms well more advanced than the typical benchmark-driven algorithms," he says, noting that he foresees more growth in the area of arrival price algorithms as opposed to the typical VWAP or TWAP algorithms.
Wald adds that time and trading in dark pools has created a more solid data foundation for algorithms to improve going forward. "Having an algorithm that knows what particular scheduled cross is taking place and when, and knowing what stocks typically trade on those crosses and on what volume they typically trade is key," he explains. "The algorithms are aware of upcoming events and can take stock away from places that aren't being executed and move to areas of higher probability, participate in the cross, and continue to work and redistribute to other areas simultaneously."
Lee Morakis, managing director, portfolio and algorithmic sales, at Merrill Lynch, says ever-improving and increasingly sophisticated dark algorithms are on his mind as well. He believes that more-specific crossing parameters as well as the ability to leverage more-sophisticated data and statistics will contribute to more-advanced offerings.
"A lot of changes will come from the amount of execution data from dark venues," Morakis says. "In talking to clients we have a better understanding of their experiences and their needs for more-advanced and customizable technology."
Morakis adds that these include more-advanced crossing order types, better anti-gaming logic and more-efficient logistics among dark pools. As dark pool liquidity obviously is important, so to is mitigating information leakage or signaling risk, he explains.