Printer Friendly
The firm has been working since the first of this year to integrate its electronic trading platforms and teams and is now leveraging that work with key enhancements to its algorithmic trading strategies.
Bank of America Merrill Lynch's suite of algorithms includes 15 core equity and six core options algorithms. Clients can also customize algorithms internally.
"If there is anything we have learned it's that in this business you have to constantly make improvements," says Michael Lynch, head of Americas Execution Services at Bank of America Merrill Lynch. "The focus on best execution is paramount for us."
Ruth Colagiuri, head of Electronic Trading Products at Bank of America Merrill Lynch, adds, "We've invested considerably and will continue to invest in improving our algorithms to ensure they are providing the highest quality execution and are customizable for our clients needs."
She explains that clients have put a heavy focus on execution quality and performance, as well as service levels.
"The market structure is continually changing and it is definitely not the case that you can put an algorithm on the shelf and say its done. We have to continually evolve our algorithms to adjust to changes in the market," says Colagiuri.
The latest enhancements include improvements to limit order placement and expanded implementation shortfall algorithms.
"Our limit order model is used by all of our algorithms to react to a market event or a trade schedule and make a decision on how to price an algo child order," explains Colagiuri. "For example, we've improved on this by introducing logic that gives a short-term signal when it is an opportune time to cross the spread. So it watches the market in real time and takes a signal that says this is a good time to be aggressive in the market."
She adds that the firm's scheduling algorithms, such as VWAP and TWAP, have typically been more passive, but this is a new shift in the balance of passive and aggressive with the enhanced logic. "It is more aggressive and with this model and we have see performance improvements of around 20 percent over a large data set of orders that we evaluated," says Colagiuri.
For the implementation shortfall algorithms the firm has introduced an enhanced market impact model. "Based on this model, the algo will schedule orders in discontinuous bursts throughout the life of the order. These improvements have made the strategy more dynamic and make the trading more granular," says Colagiuri.
Enhanced anti-gaming protection was also a focus of the algorithm improvements. "As the dark pool landscape changes we continue to build more sophisticated anti-gaming into not just our dark algorithms but our other algorithms as well," Colagiuri notes. "It includes protection from behaviors that impact the quality of execution."
The team has also addressed client requests for optimized algorithms for trading smaller orders. "We have tailored versions of these strategies for several large clients and they became popular so now we have included an off-the-shelf version that clients can use, or the option to customize themselves," she adds.
