Greenwich TCA Study Shows a Divided Buyside

Survey's surprise is that one in five firms doesn't use TCA

Despite widespread usage, the buyside has mixed feelings about transaction cost analysis and its effectiveness.

A Greenwich Associates report on transaction cost analysis, or TCA, released today outlines the benefits, shortcomings and how firms are using TCA. Greenwich based its report on a survey of 104 buyside institutions it conducted last month, between May 4 and May 9.

One unexpected finding, according to the report, shows that 19 percent of the survey sampling does not use TCA to measure its trading costs. “We were surprised to see such a large proportion of institutions not using TCA in light of its importance from an ERISA and regulatory perspective,” said John Feng, a Greenwich consultant who worked on the report.

The Greenwich study shows that 80 percent of those institutions surveyed use some form of TCA in their equity trading operations. But TCA’s ability to help generate additional returns to investors strongly divided the survey respondents.

“When it comes to the question of how important TCA is in determining investment results,” said Jay Bennett, a Greenwich consultant who also worked on the study, “35 percent say TCA plays little to no role, and an equal share say TCA is very important or extremely important.”

For the most part, the buyside sees the benefits of TCA when it comes to measuring their own desks and for compliance. But 12 percent reported that TCA added “little” value in either area.

The report noted that the need to understand and rein in trading costs has risen in recent years, as the buyside has taken greater control of its orders with electronic trading tools. These factors have put a spotlight on TCA, which is the only option right now to measure trading costs and help to better understand the trading process. Overall, the study said traders see “potential” in TCA, but it currently has some weaknesses and a need for improvement.

Traders were “generally” pleased with their TCA systems. But even here there were divisions. Exactly 43 percent of the survey said their systems were “very effective or extremely effective”; 31 percent said their systems were “moderately effective” at measuring the desk; and about 25 percent reported that their systems were “somewhat or not at all effective.”

The TCA system of choice is from a third party, according to the survey. Roughly two-thirds of the respondents use a third-party vendor. Of this group, about half use ITG and about a quarter employ Abel Noser. The balance of TCA users is comprised of those using proprietary systems, 11 percent, and those using broker analytics, 8 percent.

The survey also looked at how the buyside views TCA in relationship to broker performance. One-third of the survey reported that TCA was “extremely effective” in measuring brokers; one-third said TCA was “moderately effective”; and one-third said TCA was “somewhat or not at all effective.”

One hedge fund trader quoted in the report said that it is unfair to measure brokers with TCA and not include other factors that go into a trade. “Every trader knows you give your toughest orders to your best [sellside] traders, but these traders will inevitably be penalized if judged solely on TCA,” he said.

The survey was made up of 68 institutional money managers, 12 hedge funds and a smattering of mutual funds, pensions, endowments and banks in the U.S., Canada and Europe. Those who responded consisted largely of traders and head traders, but it also included chief investment and chief operating officers, according to the report.

Buyside participants were mostly pleased with the data delivery systems, data and consulting services of their TCA providers, the report said. Still, they compiled a list of areas that could be improved. They included:

  • TCA systems should provide faster and “real time” data
  • TCA should adjust for changing conditions across the life of an order
  • The systems should cover more asset classes
  • They need to involve data from trading sources beyond the primary markets
  • They need to do a better job of showing how outliers can skew overall results

One participant from a U.S. investment company who heads its global equity trading division told the report’s authors that he likes TCA system interfaces. But he doesn’t like how they organize data, or how long the overall TCA process can take.

“Providers need to refine their data to make results more actionable on their own,” he told them, “and to improve their consulting services and analytics to ensure that the data their systems produce lead to meaningful conclusions.”

Referred to in the study as a “striking finding,” many buyside respondents reported that they would change their trading style if TCA became a larger component of trader compensation. Of 73 respondents, 57 percent-or 42-said they would alter their strategies and practices.

“In addition,” the report noted, “some traders would be tempted to game the system to produce results that maximize compensation.”

They noted to the report’s authors that the drivers of TCA results weren’t always in line with “client goals and shareholder interests.” Traders said it wouldn’t be good to concentrate on beating benchmarks rather than efficiently implementing investment ideas.

In fact, 23 percent of the respondents said TCA should play no role in trader compensation. At the other end of the spectrum, 20 percent said it should play a “large or extremely large role.”

The report concluded by saying that the buyside would get more out of TCA if it were better integrated into the overall investment process. “Doing so will require institutions to include everyone who participates in the investment process into the TCA process,” the report said, “and to train them in how to properly interpret and act upon results.”

Stepping back from the survey, Bennett said traders are torn on a variety of issues relating to TCA. They accept the regulatory and fiduciary responsibility to demonstrate best execution. But they also reject any notion that they didn’t live up to a model’s expectation, he added.

Bennett said that it is also difficult today for a buyside shop to determine how aggressively to apply a system that doesn’t deliver “perfect information” in real-time. However, he also offered an assessment of TCA.

“If I were running a buyside trading desk, would I rather have it or not have it?” he asked. “I’d rather have it.”