|
 Can transaction cost analysis really lower my cost
of trading? Can transaction cost analysis really lower my cost of trading? Yes. Effective trading involves making a decision about a strategy to follow in getting the trade completed and continuous assessment of market conditions during the period of execution. Transaction cost analysis has little role to play in the latter, which involves real-time monitoring of market conditions and price fluctuations. However transaction cost analysis is critical in establishing the most resource efficient and cost effective execution strategies for different types and sizes of trades being completed for various portfolio managers in a wide range of securities. It may also impact the choice of executing broker. Better execution strategies will undoubtedly lead to improved execution performance and lower costs. Can analysis ever be sophisticated enough to account for all the factors in trading? No. There will always be some contextual aspects of trades that cannot be captured and adjusted for within the analytical process. However, for most trades the information that can be captured forms a sufficient basis for analysis and conclusions. This is especially the case where the number of transactions is statistically significant. In any analysis it is important to distinguish between generally applicable conclusions and the position surrounding the outcome of individual transactions. What is the most appropriate benchmark for me to use? There is no single benchmark that is appropriate for all transactions. An effective trader will review their activity against a number of different measures to better understand the impact of trading on overall investment strategy and performance. In addition traders acting for different portfolio managers different funds will need to take account of their particular trading objectives and concerns. If a single benchmark was sufficient then the process of analysis, and ultimately of trading itself, would become purely an algorithmic exercise with no need for human intervention. GSCS believes that trading is as much an art form as a science. Application of mathematical analysis may assist traders but it will never replace them. What can transaction cost analysis tell me that I do not already know? A great deal. All traders have a clear and unambiguous priority to complete the trades given to them. This makes it difficult to make time to analyse and quantify aspects of trading that are important to performance. While certain transactions tend to stand out, it is dangerous to draw general conclusions from what is essentially anecdotal evidence. Good transaction cost analysis allows traders to benefit from the analysis offered by the service provider, which will almost certainly contain insights about general patterns of trading and performance that might otherwise be lost. What actions can I take based on transaction cost analysis? There are essentially three different areas in which proper analysis of trading costs can be used to improve performance. The first of these is within the trading area itself. Analysis can lead to changes in the trading strategy followed on certain types of trade. In some cases this will simplify the trading process (e.g. using DMA or algorithms) which achieve appropriate trading results while lowering commissions and improving the efficiency of the trading desk. In other cases more complex strategies may be identified. The second area is in relation to dealings with brokers. Being able to more effectively analyse their performance, ensuring that due allowance is made for trade complexity, allows traders to ensure that brokers are given trades that match their strengths more precisely than might be the case with less thorough information. Finally by assessing the kinds of trades given by different portfolio managers and explaining the consequences in terms of effect on investment returns, traders can positively impact the overall investment performance of the firm in a constructive and positive way. How important is peer group comparison? Peer group comparison provides a useful "high level" indication of the competitiveness of execution performance and commission levels. However, the differences between companies are considerable (even where the overall investment approach and size are comparable). This makes it extremely difficult to draw meaningful detailed conclusions that are genuinely reliable. Overall greater improvement will come from more sophisticated analysis of a manager's own trading than from peer group comparison. Why should anyone other than traders care about transaction cost analysis? It is increasingly recognised that effective trading performance is an integral part of overall investment returns. In periods when average absolute returns are at historically low levels, the importance of trading becomes even clearer. Portfolio managers and senior management are therefore ever more interested in transaction costs. The same factors are driving the interest of external clients, especially pension funds who see transaction costs equal to or greater than investment management fees. Finally regulators are keen to ensure that client funds, in the form of commissions are spent appropriately and that best execution is not affected by bundling of execution and research. They, and the compliance officers who deal with them, therefore have a growing and legitimate interest in transaction costs. What is implementation shortfall and why is it important? Implementation shortfall compares execution prices against prices prevailing at the time the portfolio manager decides to trade. Any difference represents the cost to a fund of achieving its desired exposure, whether caused by delay in decision making, delay in trading or the impact of completing large orders in less liquid securities. In aggregate it represents the direct loss of investment "alpha" (i.e. return) caused by the act of completing the trade. Implementation shortfall is important because it shows how much the complete cost of trading. As such it is a useful guide of the extent to which an investment “idea” had to outperform in order simply to “break even”. This can affect how portfolio managers evaluate their activity. However, implementation shortfall does not take any account of the practicality of trading in any specific market. It is an idealized number rather than a realistic one. It is therefore essential when considering implementation shortfall to disaggregate the various components in order to properly analyse trading performance. What are the advantages/disadvantages of volume weighted average price measures? Volume weighted average price (VWAP) measures reflect all activity in the market over a period of time. The advantage is that they are easy to understand, relatively easy to measure and provide a “benchmark” that is, in effect, the performance that should occur if the trading process adds no value. Indeed many brokers have tools today, available to clients, that effectively guarantee to achieve the VWAP as an execution price. In that sense they are comparable to “index returns” as a basis for measuring investment performance. The main difficulties reflect the fact that not all trades are ones in which any given trader could “participate”. As such the VWAP numbers may need to be “adjusted” to provide a fair benchmark. The way in which this adjustment takes place can be confusing and is not always completely transparent. Finally because they represent a “default” result achievable without positive input from traders, traders as a group tend to hold them in poor regard. Should I look at average performance or focus on specific trades? Both. The impact of trading performance on investment returns is driven by the average performance over all trades and this is where any analysis should begin. One-off trades, good or bad, may be more easily recollected but they do not provide a meaningful basis for professional or sophisticated analysis. Only by properly understanding what the general causes of particular costs are, can actions be devised that will consistently reduce those costs, thereby improving returns. However, it is also important to note that averages can and do hide wide variations in performance. Therefore any consideration of actions based on average performance need to take proper statistical account of “outliers” that have a big impact on the average numbers being reviewed. In this sense individual trades can be relevant and should not be ignored. Can transaction cost analysis help me evaluate my brokers? Yes. Not all brokers are equally effective trading all kinds of orders. Agency brokers in particular may find it more difficult to be effective with very large orders. Even the large global dealers however, are not as strong in some areas as others. In determining which brokers are best suited to handle different orders it is critical that any analysis of historic trades takes account of as many of the differences in the types of order as is practical. Therefore in looking at capabilities of brokers it is necessary to consider the trading landscape within which they have been asked to operate. To the extent that these factors are properly adjusted for within the analysis, it is appropriate to use the data to compare and contrast the performance of brokers and to adjust the pattern of future trading to take account of the relative strengths and weaknesses identified. How do different market structures affect the cost of trading? Not very much. The three main kinds of market infrastructures are those with competing market makers, those based around order book that matches buyers and sellers and crossing systems. The characteristics that shape execution quality and the cost of trading tend to be liquidity and volatility. The latter tends to be determined by the type of security not the market structure, though arguably market maker structures may be more volatile in the very short term. Liquidity is not determined by market structure either. Therefore market structure does not really have any direct correlation with likely trading costs. Do clients really care about transaction costs? Yes. The very largest and most sophisticated pension funds and investors understand that costs in general, not just transaction costs, have a significant impact on long term investment returns. Costs as a whole are the second most important factor in long term returns, coming after asset allocation but ahead of investment manager selection in importance, as assessed by independent studies. Clients increasingly recognize that transaction costs, including commissions add up to a total that is comparable, and in some cases greatly exceeds, investment management fees. Transaction costs are therefore in many cases the largest cost being incurred. In such an environment clients would always be interested in transaction costs. Their interest has been further encouraged in some countries by the impact of regulators. How does the size of my transactions affect the transaction costs I incur? In general the larger the transaction, the higher will be the cost. This is particularly the case where the trade being executed is both large and is a “taker of liquidity” from the market (e.g. buying a security that is already in demand and whose price is rising). In circumstances where the execution is “providing liquidity” (i.e. selling a security that is in demand and whose price is rising) there is a commensurate reduction in cost. However there is an asymmetry between buys and sells which means that over a significant number of transactions the net effect is one that means that larger trades cost more. Should I measure the performance of traders based on transaction cost analysis? Yes. Obviously traders like any other employee need to be given clear objectives and a basis of measurement of performance that is fair. Transaction costs should be a component of their objectives. Creation of a fair measure is difficult. As was noted earlier creating a comparable landscape of activity so that the degree of difficulty of traders is properly taken into account, is not easy. This is especially the case in larger firms with more diverse investment styles and trading requirements. For this reason it is likely that any direct tie between compensation and trading costs would cause more problems than it could solve. Trading is an art as much as a science and as such evaluation of its worth should be, at least in large part, a qualitative not a quantitative process. |
 |
|
FAQs - Investment managers
