A reduction in the costs associated with the entire trading implementation process will lead to better overall fund performance, although it is not unknown for funds with low transaction costs to underperform their peers due to poor upstream execution.
One conundrum to transaction cost analysis (TCA) is that a variety of benchmarks and calculations are applied to analyse transaction costs such as Volume Weighted Average Price (VWAP). It might not always be the most appropriate metric. For instance, a small cap strategy will employ a certain metric, a large cap desk an entirely different one.
“There is a lot of buzz currently surrounding TCA as a product,” says Joseph Wald, chief executive officer of EdgeTrade Inc. in New York.
But before any TCA tools become relevant, the buy-side trader needs to address their execution objectives for a particular trade, portfolio or investment decision.
“While we know that VWAP is something that many look at, end investors have to question whether VWAP is offering them best execution. And, on the pre- and post-trade side, unless you are tracking the right variables, the results are not likely to mean anything unless you can match the appropriate inputs with the appropriate outputs,” Mr Wald adds.
Pinching the pennies
As pension funds and other institutional investors strive to earn better returns, they are clearly scrutinising ever more closely the cost of achieving that elusive outperformance.
John Wightkin, co-founder and managing director of Quantitative Services Group LLP (QSG) in Chicago, says: “Realised investment performance depends on two processes - the expected returns coming from the stock selection process and the costs of implementing the portfolio strategy.”
Historically, investment professionals have concentrated only on the stock selection process when trying to enhance their overall returns. With regulation and the backdrop of a lower return environment, pension funds and trustees have started evaluating the implementation process as a source of extra performance.
“Within this area, lower trading costs are extremely desirable. However, improvements to other parts of the implementation process – more efficient dissemination of portfolio ideas to traders and swifter trading decisions - are equally as important to lowering the overall implementation costs,” adds Mr Wightkin.
QSG is among the TCA service-providers to have witnessed an uptick in demand for its services. The company evaluates the total implementation process to ensure that it is being managed as efficiently as possible.
“We have actually seen firms with low transaction costs underperform their peers as a result of poor execution upstream in the trading process. Fundamentally, reducing all costs associated with the implementation process – which includes the transaction costs – will lead to better overall returns,” he adds. Lowest transaction costs do not necessarily imply best trade execution.
In-depth analysis
“To determine that [best execution], the measurement system must evaluate the total implementation process. Such evaluation includes examining the quality of the actual executions and determining the value and impact of the stock research and selection process,” contends Mr Wightkin.
Duncan Begg, head of analytical products and research products ITG (Europe) in London, adds: “Certainly paying lowest commission does not guarantee that one is going to achieve the best trading result. While you could end up paying minimal basis points in brokerage, a disproportionate amount of cost in the implementation of the order may be incurred.”
Having an objective model for assessing an expected implementation cost for an order is critical.
If in a hypothetical example, trader ‘A’ deals in small caps, and trader ‘B’ handles large caps, on a straight implementation shortfall basis A will always underperform B. But is it necessarily right to say B is the better trader? Trader A may be the better trader once the difficulties of the orders they are trying to execute are factored in. “Naïve implementation shortfall is therefore a poor way to assess whether trading costs are ‘reasonable’,” Mr Begg argues.
QSG’s transaction cost analysis starts with the nightly delivery of its client’s trading data through an order management system (OMS) or the client’s custodian.
Mr Wightkin explains: “With this data we perform several different levels of analysis to arrive at a comprehensive review of a firm’s execution process. Such analysis includes the standard benchmark analysis, a detailed attribution analysis (to arrive at the factors driving the trading costs), and an execution quality analysis.”
ITG’s Mr Duncan adds: “It’s absolutely key to assess the most appropriate benchmarks. It is also important to look at the slippage that occurs at the various stages in the life cycle of an order and to conduct analyses through time and cross sectionally (eg between orders of different size), to answer whether there is a systemic influence impacting the trading result.”
From Mr Wightkin’s North American perspective, pension funds are increasingly interested in understanding their managers’ execution processes.
“By measuring their managers’ trading costs, they can better analyse their manager’s execution effectiveness and ensure that they are receiving ‘best execution’. They are also utilising the information as an additional input into their evaluation of their manager’s performance.”
He adds: “Combined with the return attribution and analysis, these pension funds now have more comprehensive information to judge and compare their managers. In the end, ultimately the pension funds want a better picture of their manager’s overall investment process.
Rick Di Mascio, chief executive and founder of London-based Inalytics, says pension funds want to know how much transaction costs impact performance and return, as well as how efficient their fund managers are. (The firm recently launched a new service - Identification of Manager Skill, which as a product and even a concept hardly existed a a year ago).
For Inalytics, which has a client base split equally between pension funds and fund managers, equity trade and transactional data is accessed from the pension funds’ custodians, while asset managers and brokers send data to them from their OMS – usually on a quarterly basis. Regulators too are clients.
Breaking it down
Each trade is then analysed in terms of costs to establish the overall impact of transactions on performance. While tax and commission are what people usually associate with transaction costs as they appear on the contract note, 70 per cent of total transaction costs are accounted for by ‘price slippage’ in executing trades - from price moves between the decision to send a trade order and actual execution.
Mr Di Mascio says: “The quality of the data has certainly increased significantly over the last few years. When we first started out eight years ago, data quality was a real problem, however, this is much less of an issue now. When we analyse the costs now, we are able to say if they were incurred at a reasonable level or not and if they were above/below benchmark.”
“The second challenge in analysing equity manager’s trading costs is the interpretation of the data. If the analysis does not incorporate the investment style underlying the execution process, any interpretation may not be entirely relevant to the manager being analysed. The trading costs have to be put into the context of the overall investment process to truly have meaning in the overall analysis,” says QSG’s Mr Wightkin.
TCA is predominantly confined to equities today. Models and calculations have hardly been applied to other asset classes such as foreign exchange and fixed income. Many practitioners say that it is more difficult to conduct a transaction cost analysis of a bond portfolio due to its poor data transparency and putting together benchmark data.
Mr Wightkin says: “Most clients are looking to measure their equity executions, which may largely be because equities are furthest along in measuring and analysing trading costs,” QSG is nevertheless “beginning to receive” questions about measuring fixed income, FX and derivatives (options and futures).
It’s unlikely to be a VWAP solution in fixed income arena as it would be extremely hard to span all the esoteric, illiquid and infrequently traded bonds. Illustrating the problem, Reuters DataScope Select was launched this January with a team of evaluators
pricing illiquid fixed-income instruments covering one million US securities as well 60,000 European securities in four major currencies.
Measuring shortfall transaction cost analysis with fixed-income could be done by comparing indicative received bids prior to a trade being executed and comparing the expected traded price with the actual traded cost. But according to Lisa Manuele, head of BNY Global Transition Management at BNY Brokerage, one cannot compare the data and so “it would not make any sense.”
Much the same carries with currency, although the 11am New York City time fixing has provided a standard measure for FX execution analysis. Still FX is opaque and not centrally traded.
One firm, Hotspot FXi, which developed the first liquid, client-to-client and bank-to-client ECN Internet foreign exchange trading site, amalgamates prices from 10 market makers. An institution executing, say, $500m (€420m) a day in FX, could incur $12.5m of transaction execution cost per annum if even one basis point was charged to it - the smallest unit in the FX marketplace. But some client studies have indicated that these charges can be as high as 70 basis points, depending on the execution approach.
The biggest issue with measuring and analysing trading costs in the fixed income, currencies and derivative is the access to detailed and timely data. Without this data, the analysis has to make assumptions concerning the measurements. These assumptions can lead also to questionable analysis and misleading conclusions, Mr Wightkin contends.
Clean and rich
ITG’s Mr Begg adds: “Being able to provide clean and rich data from the order management system (OMS) is key. This in turn suggests a level of sophistication is required of the OMS.” ITG, which trademarked ‘TCA’ as a brand, believes that “amassing a large quantity of data, use of a powerful database and assessing the most appropriate benchmarks” are all part of a sound evaluation process.
Some think that with so many broker-sponsored TCA products available, the buy-side should exercise scepticism over who is actually offering these services.
“At the end of the day, a TCA product should be a holistic solution that encompasses for a buy-side client all of their executing brokers – not something from a single specific broker – possibly along the lines an independent transaction cost analysis provider,” argues EdgeTrade’s Mr Wald.
