Bad performance numbers and regulatory pressure are forcing buyside desks to pay more attention to trading costs. Desks that are looking beyond brokers' commissions are building systems or are subscribing to third party offerings. The goal is to measure their implicit – timing and opportunity – costs. Armed with better information, they can then revise their trading strategies.
Among the top providers of transaction cost analysis are Plexus Group, which was acquired by J.P. Morgan Chase this summer, Elkins & McSherry, a unit of State Street, Instinet, ITG, Abel Noser and BECS.
Now a relative newcomer is stirring things up. Quantitative Services Group, a suburban Chicago consultancy, markets a transaction cost measurement system called T-Cost Pro that lets traders evaluate their executions one day after they occur.
The six-person firm was founded in 2000 by former Donaldson, Lufkin & Jenrette quants Tim Sargent and John Wightkin, following their employer's acquisition by CS First Boston.
Both execs have several years of quantitative trading and research under their belts. Sargent began his career at Northern Trust with responsibilities for index funds. He later had stints at Merrill Lynch and Salomon Brothers before establishing a global quantitative research effort at DLJ.
Wightkin also started on the buyside at Ohio Teachers as a quantitative analyst and head of the trading desk. He later moved to Weiss Peck & Greer and a research boutique called Chicago Analytics before joining DLJ.
Both execs discussed the T-Cost Pro service with Traders Magazine technology editor Peter Chapman.
Traders: Why is transaction cost measurement hot right now?
Sargent: There are a few different reasons. First, of course, is the overall low return environment. With returns negative or modest at best, transaction costs play a larger role. The other would be the regulatory climate.
Traders: Why. New rules coming out of Washington?
Wightkin: No. But the Securities and Exchange Commission is stepping up its auditing and trying to develop a best practices document. Also, AIMR [Association for Investment Management and Research] is coming out with trade management guidelines that will provide some structure as to how to measure best execution. One of their recommendations is to quantify and measure trading costs and the quality of trading. AIMR is emphasizing the need to have some transaction cost analysis in place.
Traders: O.K. How does T-Cost Pro help?
Sargent: T-Cost Pro is a forensics device. We developed a new set of measurement techniques to help the decision-making process. Rather than rely on traditional benchmark analysis, we wanted to come up with a methodology that would give us a measure that traders and portfolio managers could use to make better decisions.
Traders: What's unique about T-Cost Pro?
Sargent: Much. We can give you a few examples.
Traders: O.K.
Sargent: First of all we look at the impact of the stock selection style on trading costs. How stock selection influences trading costs. It's much easier for us to analyze trading decisions in the context of the style that created the trade in the first place. To be successful at this you must know the origin of why you're trading in the first place. Often, that's overlooked. Once you are armed with that you can work on reducing the costs associated with your style decisions. The way you set up a pattern of decisions regarding your style and implementation of your style -that has features you can control.
Traders: Stock selection styles generally come down to growth or value.
Wightkin: We've also identified such styles as relative value, deep value, historical growth, price momentum and earnings momentum.
Sargent: We can see where the cost came from, especially when you're trading a portfolio of stocks. Traders may like to know what the influences were. Whether they're from, say, historical growth characteristics as opposed to value characteristics.
Traders: O.K.
Sargent: For example, let's say you've traded 100 stocks for the day. And the majority of your costs stem from timing consequences. You can look into the system and see that a majority of the timing consequences that influenced the whole portfolio were driven by the stocks that fell into the historical growth style. So, if you trade historical growth stocks in the future it may allow you additional insight.
Traders: Can you be more specific?
Wightkin: Let's say associated with your stock selection strategy is some short-term alpha or momentum. And your trading strategy involves trading VWAP throughout the day. So, you are parceling out orders and participating in the market. Through our techniques you can determine what your impact is on each of those parceling activities. And also how your timing decisions are impacting the overall cost of the trade.
Traders: O.K.
Wightkin: If your style calls for you to build in this short-term alpha or short-term momentum into your stock selection strategy you may have a lot of timing impact but little market impact associated with your trade.
Traders: Timing impact being the cost associated with feeding the order into the market in pieces over a relatively long stretch of time and market impact being the cost incurred by dumping a large block all at once?
Wightkin: Right. So, what happens is the alpha in your stock selection strategy is getting lost or you're giving it away to the market. By zeroing in on the style associated with the stock selection process, we might suggest you modify your execution strategy by trading earlier. This will wring some liquidity premium from the market. Thereby reducing those timing costs. You capture more of the alpha in your performance rather than giving it away to the market.
Traders: O.K.
Sargent: Because decimalization has led to reductions in the size available at the prevailing bid and offer, institutional traders have gone to splitting their orders into smaller pieces. So, as they lengthen those execution times they introduce risk. The question is: which is costlier? The demand for liquidity by running in and trading in larger lot sizes? Or is it the price movement you suffer when you wait as you parcel out the orders? We call it the trader's dilemma.
Traders: What else?
Sargent: A big difference is our ability to look at the cumulative footprint of a trade.
Traders: Cumulative footprint?
Sargent: Let's say you are a major trader in a small-cap stock. You buy 1,000 shares and end up pushing the price up a few pennies. Then you go back in again. You are trading aggressively. To get any trade done you must continually mark up the stock. Every time you buy you mark it up a little more. If you are trading in 1,000- and 10,000-share increments and you have a large order of one million shares, you could have a substantial influence as to where that stock goes.
Traders: O.K.
Sargent: So, your overall cost is not just the average cost of any one trade. It's the cumulative impact you had on that stock. We isolate that in our process and give the client a unique insight.
Traders: Anything else?
Sargent: We've constructed our own unique elements of costs. "Liquidity charges" relate to the trader's decision to trade now. "Timing consequences" relate to his trading later. We separate timing consequences from liquidity charges. Timing consequences represent the cumulative impact of the market timing decision itself separate from the charge for liquidity.
Traders: And liquidity charges are?
Sargent: The Liquidity charges are the cost of getting the trade done, often referred to as market impact. To measure these charges, most people just default to the bid/offer spread. But we break this charge into "Spread charges" the cost of trading within prevailing market bid/offer spreads and "Liquidity premiums" the cost of requesting more liquidity than is available in the prevailing market. We also review the cumulative impact of these charges, much like the timing consequences.
Traders: So, you break things down that aren't traditionally broken down?
Wightkin: Yes, we break them down into factors that have relevance to the trader or portfolio manager. We put it into terms people can utilize and make sense of. If you're going to the trouble of measuring these costs it shouldn't be just to satisfy a regulatory requirement. You should be able to get something out of it.
Traders: Clients access your system over the Internet?
Sargent: Correct.
Traders: They deliver their trade data to you at the end of the day and then you deliver a report to them the next day?
Sargent: Yes, but it's not just a report. They go into their area and review their trades for that day. It's a system that allows them to do a great deal of sorting and review in detail.
Traders: How much data do they provide at the end of the day?
Sargent: Many of our clients deliver each execution.
Wightkin: We create analysis from each of those time-stamped executions.
Traders: Data comes from client order management systems?
Sargent: Maybe through their OMS or directly from their broker.
Traders: And I assume you get a feed of every trade in Nasdaq and New York listed securities daily?
Sargent: We subscribe to multiple tick-by-tick databases
Traders: You also offer a less expensive service called T-Cost Lite. How is it different from Pro?
Wightkin: T-Cost Lite is lighter than T-Cost Pro. It will only do benchmark analysis. That is similar to what most other folks provide. It won't go into detail because we won't have access to each execution. T-Cost Pro is measuring each transaction and breaking into these components that make sense from a trading decision standpoint.
Traders: T-Cost Lite uses average price data while Pro analyses every trade?
Sargent: Right. When people trade several lots over the day they may not recover every execution. They get an average [price] from their broker. So, if you only have the average price data, you can't do the detailed kind of analysis. The average price incorporates all the executions but doesn't show you each one.
Traders: T-Cost Pro costs more?
Sargent: Yes. Because it examines the impact of each trade it adds more value.
Traders: Does you system allow traders to compare their performance against the popular VWAP benchmark?
Wightkin: Yes. In T-Cost Pro, we provide measurement against 15 standard benchmarks including VWAP.
Traders: Can they evaluate their brokers as well?
Sargent: We do break transactions down by brokers in both products.
Traders: You've entered into an arrangement with vendor Eze Castle to integrate your technology with its OMS Traders Console. Why? What is the set-up?
Wightkin: We do, of course, work with other OMS providers, but this unique arrangement with Eze was client motivated. Their clients get this integrated product. That allows them to evaluate what their trading strategies should or could be. They can now do their trading and post trade analysis all on one box without having to worry how these two talk to each other.
Traders: Are you affiliated with or owned by any broker dealers?
Sargent: No.
Traders: Finally, the acquisition of Plexus by J.P. Morgan Chase was said to have been driven by Chase's role as custodian in transition management transactions. Is transition management an area in which QSG gets involved?
Sargent: Actually, we just completed a project where we built custom software for a major custodian to provide that service. We don't provide that service directly, but we think it's a huge business. And one that deserves a lot of attention from the measurement standpoint.
Traders: Thank you, guys.
