Transaction Cost Analysis A-Z

Transaction Cost Analysis A-Z — November 2008

V. Trading Performance Measurement

or contrarian strategy. For example, the opening price delivers performance measures that are easily biased. Momentum investors will push down the trading performance measure because they buy (sell) when prices have risen (fallen), so the opening price is low (high). By contrast, contrarian investors buy (sell) when prices have fallen (risen), so the opening price is high (low). This pushes up the trading performance measure. Benchmark comparison may be gamed Gami ng p r o b l ems a r i s e wh e n intermediaries know the benchmark that will be used to measure their execution performance and consequently time their trades based on their evolving benchmark score. The consideration of trade-timing is important for investors who give their intermediaries discretion over the timing of their trades. Proper trading performance measures should help investors ensure that their brokers make appropriate use of this discretion. Trade-timing effects are best measured when the benchmark does not depend on the time of the trade. When the benchmark relates to the execution time, intermediaries can accelerate/decelerate trades based on their evolving benchmark score. This behaviour is questionable because it can result in higher total transaction costs for the end investor. Harris (2003) describes several situations in which brokers can game their evaluations. For example, brokers who have discretion over how aggressively they fill orders can easily game a performance measure based on the spread midpoint prevailing at the time of the trade. They will simply supply

liquidity and never take it. Hence, they always buy at the “bid” or sell at the “ask”. This behaviour makes their performance look great but comes at the expense of increased timing risk. Benchmark comparison needs a unique and relevant reference price Benchmark comparison assumes that the benchmark is an appropriate price for the value of the security. The only “true” value of a security is the price at which an actual trade was made. Sometimes, the benchmark is not such a price. For example, the spread midpoint, the VWAP or the LHOC may be prices at which no trade has taken place. Furthermore, market fragmentation and proliferation of liquidity pools make the determination of the right benchmark more difficult. Indeed, the benchmark comparison becomes conceptually difficult in the absence of a recognised unique price for the securities traded. When a security is traded at the same time on various execution venues, which of the coexisting prices is the benchmark price for the security? With MiFID around the corner, this issue is of great importance. By allowing multiple trading venues to compete, this new piece of European regulation is, by design, allowing liquidity pools to fragment and putting the very existence of recognised benchmarks in danger.

Benchmark comparison is not standardised

Finally, for at least two reasons, benchmark comparison does not offer a unified framework to enable easy assessment of execution performance across a series of trades at any aggregate level.

60

An EDHEC Risk and Asset Management Research Centre Publication