Trading algorithms have long been used for large-cap stocks, but historically, they have not been so adept at trading smaller, less liquid names. That, however, has been changing.
Over the past several years, the sellside has introduced a number of new algos that target small caps and other illiquid equities. (See April’s Cover Story.) Though they use a variety of strategies, they all have a common goal: get the trade done without information leakage moving the stock price.
In 2009, Knight Capital unveiled a new algorithm specifically designed for small- and mid-cap stocks. Dubbed Oasis, the system promised to find liquidity in smaller names amidst what seemed like a desert in the post-crash environment.
According to Joseph Wald, managing director at Knight, small-cap traders are worried at least as much about getting size done as getting the best price in an individual trade. The more small trades that get done while the bulk of the order remains unfilled, the more chance there will be information leakage adversely affecting the remaining trades.
“We developed technology that directs orders to maybe two or three venues getting size out of those venues, while minimizing information leakage, versus having to go out and ping 30 venues and potentially not getting what we want,” Wald said.
And the strategy is growing. In 2010, Oasis traded a notional value of $11 billion in stock. In 2011, that number was $25 billion.
Knight isn’t the only firm worried about getting size done as quickly as possible. Todd Lopez, head of sales for equities electronic trading in the Americas at Goldman Sachs, said one method his firm uses to capture size is to utilize contingent orders.
Algos can place large contingent orders in one or two venues while at the same time sending multiple small orders to a number of different venues. If the entire large order can be filled at once, the small child orders are canceled. If some of the smaller orders have already been filled, the contingent order will execute at a smaller amount. That way, there’s zero chance of overfill, Lopez said.
Some other algos use similar methods, and Lopez sees the strategy as something the industry as a whole is starting to embrace in an attempt to find blocks.
