A data-driven broker selection tool demonstrates a process-driven, quantitative approach to execution performance measurement.
As the implementation date for Markets in Financial Instruments Directive (MiFID) II fast approaches, regulators are sending a clear message to the investment community: Make sure that you perform robust analysis of the effectiveness of your order execution arrangements.
This should come as no surprise. As part of the FCA’s 2014 Thematic Review TR14/13, the U.K. regulator expressed its withering assessment that “most firms lacked effective monitoring capability to identify best execution failures or poor client outcomes,” a requirement of MiFID II. The FCA went on to demand that “firms must take action to ensure that their monitoring is helping to deliver best execution for clients on a consistent basis.”
In short, the requirement to implement regular, effective monitoring is clear and unambiguous. The more pressing question, then, becomes one of approach—how to ensure that the monitoring process deployed will meet MiFID II’s standard of taking “all sufficient steps to obtain … the best possible result for their clients.”
To address these new higher standards, an acceptable monitoring process should combine at least the following aspects:
- It should employ data-driven quantitative approaches
- The data should be sufficiently clean and structured in a way that readily allows for analysis and consolidation
- The data collected should be unbiased, allowing for direct comparisons of execution quality
- The monitoring process should be ongoing and repeated regularly
Using a data-driven selection tool
One approach to helping meet the monitoring challenge involves decoupling the selection of an execution strategy from the selection of a broker. Under this method, a buy-side trader would retain focus on strategy selection (VWAP, VP, IS, etc.) while the choice of execution broker is determined by preconfigured allocation percentages.
Because the allocation a broker receives is driven by the broker’s target percentage of orders (and not other considerations), the execution sample should be less prone to bias in its composition. Assuming enough flow is sent, all brokers should receive a representative sample of orders spanning different liquidity profiles, volatility groups, spreads, industry sectors and countries. This would provide a better foundation for comparing execution quality among brokers.
A data-driven broker selection tool is designed to be the infrastructure component that manages the transmission of orders to execution brokers in line with allocation targets. It consists of a database of historical order allocations and fills, a network providing connectivity to execution brokers, and the algorithm for determining to which broker a new order will be sent. Additionally, by serving as a centralised hub through which order executions flow, this tool naturally functions as a repository for collecting the resulting execution data in a standardised format.
In the context of ITG’s data-driven broker selection tool, Algo Wheel, the allocation algorithm is driven exclusively by percentage targets explicitly set by the investment manager (e.g., 10% to Broker A, 8% to Broker B, etc.) These targets can be updated at any time, and it is up to investment managers to decide how often they wish to analyse execution performance and potentially update broker allocations within the algorithm.
A data-driven broker selection tool is not necessarily touted as a turnkey solution to automate and analyse all execution performed by an investment manager. Different orders require different levels of attention. In the case of illiquid securities or difficult markets, taking “all sufficient steps” may require a much more hands-on approach to order execution.
Instead, the selection tool is commonly used to assess the quality of execution for orders that do not require the careful attention of the buy-side trader. Depending on factors such as portfolio concentration, institutional versus retail client base, investment style and inflow/redemption activity, the proportion of orders that fit the mold for a selection tool will vary.
For some managers, appropriate orders will account for a small proportion of orders and the tool may be of limited benefit. However, for managers whose daily flows significantly feature uncomplicated orders, the broker selection tool potentially delivers better workflow efficiency and execution monitoring.
Moreover, its implementation demonstrates a process-driven, quantitative approach to execution performance measurement. Judging from the language surrounding MiFID II, that is something the industry can expect regulators to focus on beginning 3 January, 2018.
We’d love to hear your feedback on this article. Please click here