Coping with fragmented liquidity and algorithmic order flow remains a challenge for buy-side firms. According to Matthew Celebuski, senior managing director and head of the quantitative research and trading team at SJ Levinson & Sons, improved transaction cost analysis tools -- such as the boutique institutional brokerage's recently launched quantitative analytics platform, Trade Analysis Program (TAP) -- can help buy-side traders and portfolio managers maximize alpha. In an exclusive interview with Advanced Trading, Celebuski explains why the buy side would be significantly disadvantaged without pre- and post-trade analytics as well as concurrent analysis of real-time trades.
How are trade analysis tools evolving to meet the buy-side trader's needs in a fragmented, high-speed market?
Celebuski: There are two ways analysis tools can help: one is scheduling, and the other is execution efficiency. When traders get an order, they have to determine how many shares to execute in what timeframe and how to interact with existing liquidity. ... The trader needs a quantitative tool to help analyze that, and this is typically TCA. This will show how these orders have traded in this sector and help the manager develop a game plan for the sector. Measuring historical performance helps traders examine and improve their process.
With algorithmic trading creating hundreds of child orders, has it become more challenging to measure trading costs?
Celebuski: Given that the market is 80 percent algorithmic in nature, ... we've gone from the shotgun blasts of liquidity events, which were negotiated among the block desks on Wall Street, to this Uzi-fire of a liquidity handshake in 100-share lots. If I told you that 80 percent of IBM traded at 47th and Lexington, you'd go there to trade IBM. Well, 80 percent of IBM is traded via computers, so that's where the liquidity is in the market. You can't trade without these tools or you'll be severely disadvantaged.
How has technology improved the ability to capture and analyze trade data?
Celebuski: We're able to capture and record substantial amounts of trade and tick data with the latest technology. The decision-process analysis also has improved substantially.
The trading systems themselves capture a higher granularity, so you can look at each fill. However, we [still] perform a high degree of data validation and interpretation due to systematic errors in order management systems.
Is there a role for pre-trade analytics, or is the trend toward real-time analytics at the point of trade?
Celebuski: Concurrent trade analysis has improved substantially, and pre-trade analysis has also improved. The ideal would be to run the pre-trade analysis on the unexecuted portion, run the post-trade on the previously executed order and a concurrent trade analysis on the working orders.
Is there a downside to using TCA to adjust a trading strategy in real time?
Celebuski: People have the tendency to hyper-focus on what is happening and lose the big picture. ... The goal is not to micromanage every order but to maximize the return to investors. ... I believe that 80 percent of the gain that you get from TCA is to find a strategy that matches the trading style of the portfolio managers and consistently improves the returns of the fund. Twenty percent of the gain is from algo selection, broker selection and trying to access the liquidity in the market in the most efficient way that is possible.
Where do you think TCA is headed in the next two to five years?
Celebuski: As the market becomes more electronic and more fragmented, No.1, you're going to see the impact of the strategy and strategic implementation in execution move to the forefront on the desk, and No. 2, the portfolio managers will take into consideration their transaction costs, which we see today with a lot of quantitative funds. ... The fundamental portfolio managers will focus much more on their expected transaction costs than perhaps they have in the past.
Will trade analysis platforms become more predictive of which strategy or algo the buy side should use?
Celebuski: TCA will be able to tell you on a post-trade basis whether you had a good strategy, whether it worked and where you executed well. If you think of TCA as that quantitative history -- and it can be minute by minute -- with an understanding that it can have some predictive power, then yes, it's going to help a lot.