Increased scale and complexity of market structures demand paying closer attention to the execution of a trading idea with the goal of preserving expected return of the trade. Professional traders need to take into consideration the important concepts of best execution and execution performance.
Best execution is an objective that goes beyond achieving the best possible trading price. Depending on the underlying trade or investment idea, best execution includes taking all reasonable steps to achieve the best possible outcome along several dimensions. These dimensions include:
o Price (execution, price improvement, spread capture);
o Cost (explicit, market impact, adverse selection, opportunity); and
o Liquidity, volatility, probability of execution, speed, and latency.
To achieve the best possible outcome, the execution process should be optimized on various levels in the investment process: portfolio construction, buy-side trading desk, execution algorithm selection, and general trading infrastructure.
For example, when constructing a portfolio, one of the parameters that measures projected returns of the portfolio/position should be the expected execution costs. If the projected execution cost is larger than the underlying alpha of the trade, a portfolio manager should consider reducing the position. Once the position arrives to the buy-side trading desk for execution, the position is then designated to the optimal execution channel (e.g., manual execution, cross, broker selection, and algorithmic execution) based on its historical trading properties and its current market conditions (i.e., size, available liquidity, and price dynamics). The position is monitored for its execution quality and market changes. The general trading infrastructure should assure adequate connectivity, system stability, and risk-handling capabilities for company-specific and broader market-related events.
KEY CRITERIA
There are a number of common execution performance criteria that the professional trader should keep in mind when selecting the appropriate execution strategy. The most general and relevant criteria used to measure execution performance is Implementation Shortfall (IS). This approach has become an industry standard as it captures the difference between the price that an investor decides to trade and the average execution price that is actually achieved. In practice, IS is calculated as the difference between the position’s arrival price, or mid-quote, at the time of arrival to the market and its average execution price.
At the level of algorithmic strategy, best execution is achieved by balancing multiple conflicting goals such as best trade price, minimal market impact, optimal time and liquidity allocation, and highest possible completion rate. We can divide them into parent and child order execution quality metrics. Parent order execution quality analytics focus on the performance of the overall trading position, such as the deviation from the desired price benchmark (i.e., open, close, arrival price, Volume-Weighted Average Price (VWAP), and previous close); market impact; price reversion after trade completion; realized volume consumption; completion rate; or opportunity cost. On a child order level, performance criteria may include analytics such as bid-ask spread “capture” (e.g., passive orders buying at the bid), price improvement (e.g., attaining a lower ask price for market buy orders), or execution/cancellation rate.
AN EXAMPLE
Let’s look at the example: A Transaction Cost Analysis (TCA) report shows that IS for a buy position you executed with the VWAP strategy is -5.1 bps, meaning that the overall execution price was 5.1 worse (higher) than the arrival price (price at the start of the trade).
A shortfall is made up of trend, impact, and trading alpha. Since the market trended up immediately after the start of trading, this may explain relatively large negative value. However, the strategy was also bought at a price one basis point higher than the VWAP benchmark. Choosing the VWAP algorithm was based on the reasoning that a trader wanted to spread the trade over some time and achieve guaranteed completion.
The trader expected that the market was likely to trend up so the trader instructed the strategy to be more aggressive than usual by shortening typical trading horizon. An aggressive VWAP algorithm is typically more likely to pay the spread when trading rather than trying to capture the bid-ask spread to keep up with the schedule. Further analysis shows the overall trend up across the duration of the order, with periodic choppiness but general uptrend for the day.
The volume profile on the day of the order was comparable with other days, so the VWAP algorithm was not disadvantaged by any unusual volume on the day. It participated fairly evenly across the entire duration according to the expected volume profile and was never more than 2% of the traded volume. That one bps cost compared to the actual VWAP algorithm could largely be attributed to crossing the spread to keep up with the schedule.
A less aggressive algorithm would have sat back more and tried to capture the spread rather than paying the spread. While trending component of the IS may not have changed given the market conditions, the overall IS performance could have improved by the proportion attributed to the market impact, which is in part driven by how often the strategy pays the spread. While 5 bps on the 500,000 notional position may not seem much (e.g., $250), these amounts may add up to a nontrivial amount over time and across the entire investment portfolio. Specifically, a selection of a “better” execution strategy could reduce average IS by half, resulting in savings of $250,000 for a portfolio with a $1B trading turnover.
QUANTIFY EXECUTION PERFORMANCE
The idea of being able to quantify execution performance is not new, and the tools to provide such quantification are now becoming common in all asset classes.
The challenge is to incorporate them into day-to-day investment and execution processes to make meaningful long-term and real-time adjustments to trading decisions.
= = =
Related Reading:
A Trader’s Survival Guide In Our Algo, HFT World by Peter Brandt
