It can be difficult for trading algorithms to capture the best price in exchange-traded funds, but now Bank of America Merrill Lynch is laying the groundwork to expand the number of funds traded by its ETF-specific algo.
The bulge firm first launched the algorithm, known as ETF-aXe, at the end of 2010. The algo leverages the expertise and liquidity provided by BofA Merrill’s high-touch desk to get best execution for more than 300 different ETFs tracking equities in the United States. That number is growing, and in coming months could include funds that go beyond just U.S. stocks.
Dan Nachtman, Americas head for algorithm product management at BofA Merrill, said the firm is increasing the number of ETFs the algo trades and is looking to cover some multi-country ETFs in the future.
“We certainly see usage across a broad spectrum,” Nachtman said. “We respond to our clients when they request we cover a certain asset grouping or a specific ETF.”
Unlike most algos, ETF-aXe looks not just at the price and liquidity of the name it’s trading, but also at the price and liquidity of the assets underlying that ETF. It also looks at futures that correspond to the performance of the ETF, if there are any.
By assessing the liquidity and price of the ETF, its underlying stocks and any futures, the algo can come up with the most efficient way to deliver the entire order. BofA Merrill then executes the trade in ETF shares for the client, while its high-touch desk takes care of obtaining liquidity in the most cost-effective manner. That might include the creation or redemption of shares of the ETF.
“This optimization and execution happens in seconds,” Nachtman said. “The client is returned only executions in the ETF, with our desk being able to assume and manage those underlying constituents, or the underlying futures.”
According to Nachtman, the ETF is often able to deliver a better price or a larger fill than a client might normally receive. Though other firms have attempted to create algos for ETFs, they have had trouble being successful, he said.
Part of the reason for that is it requires a great deal of infrastructure to take in and analyze quotes, not just for one security, but for all of the securities that underlie it. The other reason is that it also takes an experienced high-touch desk to source liquidity once the algo determines the liquidity is there.
Andrew Ruckh , head of ETF trading at BofA Merrill, heads the desk that works hand-in-hand with ETF-aXe. He said while BofA Merrill provides execution for all sorts of ETFs–including fixed-income, commodities and foreign exchange funds–the algo is currently focused on U.S. equities funds because those are the ones with the underlying assets easiest to price and trade.
An ETF tracking Chinese stocks, for instance, would be inherently problematic for an algo like ETF-aXe. If the fund were listed on a U.S. exchange, its underlying assets would not be trading at the same time as the ETF. Funds tracking foreign stocks in the Western hemisphere are in the pipeline for the algo, though. And in a few months other types of funds could be added as well.
Until then, Ruckh is focused on making sure the algo delivers optimal performance for the ETFs it does trade.
“We look at it as an upside-heavy algorithm,” said Ruckh. “Ultimately the client is going to get more shares at that same price.”
