SFC Report on Thematic Review of Best Execution

1 of 15 Tel: (852) 2231 1222 Fax: (852) 2284 4660 Website: www.sfc.hk Report on the Thematic Review of Best Execution 30 January 2018

Table of contents Introduction 1 Findings I. Governance and management supervision 3 II. Best execution factors 6 III. Applicability of best execution 7 IV. Responsibilities of execution staff 9 V. Controls and monitoring 10 VI. Arrangements with affiliates, connected parties and third parties 12

1 Introduction

1. This report summarises the key observations identified in the Securities and Futures Commission’s (SFC) thematic review on best execution. The review aimed to assess the effectiveness and adequacy of the systems and controls implemented by licensed corporations (LCs) to deliver best execution as required under the Code of Conduct1 as well as to gather views on the latest market developments.
2. Delivering best execution is fundamental to market integrity and protection of investors who rely on LCs to act in their best interests during the execution process.
3. LCs should have in place arrangements to obtain the best available terms in respect of all types of financial instruments when executing client orders, taking into account price, cost, speed and likelihood of execution, speed and likelihood of settlement, size and nature of the order and any other relevant considerations. They should also take into account the characteristics of the financial instruments and the complexity and scale of their own operations.
4. The arrangements should be subject to periodic review to ensure best execution is delivered consistently. They should include controls, monitoring and management supervision to protect clients’ best interests and ensure their fair treatment as well as to minimise conflicts of interest.
5. The review, covering 21 LCs including global financial institutions, asset management firms and local brokers, employed the following methodology. a. The SFC sent questionnaires to the LCs to understand their practices in relation to best execution, including governance structures, systems adequacy, controls and monitoring, information for clients and record keeping. b. This was followed by discussions with management and control functions of LCs to obtain a high-level understanding of the LC’s standards and practices. c. Onsite inspections were then conducted at selected LCs to assess their controls and their effectiveness in delivering best execution.
6. During the review, the SFC observed inadequacies or deficiencies in respect of: a. Governance and management supervision b. Best execution factors c. Applicability of best execution d. Responsibilities of execution staff e. Controls and monitoring f. Arrangements with affiliates, connected parties and third parties 1 Paragraph 3.2 of both the Fund Manager Code of Conduct and the Code of Conduct for Persons Licensed by or Registered with the Securities and Futures Commission (Code of Conduct).

2 7. This report highlights some good practices which go beyond our expected standards. It also outlines some less satisfactory practices which do not meet our expected standards. The examples of good practices are non-exhaustive and LCs should not regard them as the only methods for complying with regulatory requirements. 8. The SFC will keep abreast of both local and global developments in the delivery of best execution and propose regulatory changes where appropriate to align with them.

3 Findings I. Governance and management supervision Expected standards Sufficient management oversight should be in place to ensure that trade exceptions and other matters related to best execution are brought to management’s attention for timely review. Policies and procedures regarding best execution should be established to cover different types of financial instruments, including both listed and over-the-counter (OTC) products, and should be reviewed and updated on a regular basis. At a minimum, these policies and procedures should address the following areas:  factors to be considered in delivering best execution;  applicability of best execution and carve outs;  monitoring and control mechanisms to review execution quality of trades; and  the respective roles of the operational and control functions in ensuring best execution. Where applicable, these policies and procedures should also address:  handling clients’ orders in cases where multiple quotes exist and where pricing information is insufficient or quotes are absent; and  disclosure to clients of best execution arrangements including carve outs and the exclusive use of affiliates, connected parties and third parties. All relevant staff should be provided with periodic training on best execution as well as updates on internal policies and regulatory and technological developments. Observations Management oversight

1. Some LCs had a designated person or group responsible for best execution. In general, they agreed that effective management oversight is fundamental in overseeing best execution arrangements to protect clients’ best interests and ensure their fair treatment and to minimise conflicts of interest.
2. We observed the following practices which deviated from the expected standards.  Although some LCs had a designated person or group responsible for best execution, it mainly focused on equities and did not cover other areas such as fixed income. One LC set up a best execution committee but it did not meet regularly.  Some LCs frequently handled sizeable orders placed by institutional clients. However, they did not have specific reports related to best execution for management to review, eg, trade exception reports capturing execution delay and trade performance reports comparing execution outcomes with benchmarks. There was no evidence that management had any oversight of best execution.  Some LCs did not have policies and procedures requiring staff to escalate best execution-related matters to management.

4 Good practices One LC demonstrated that it had best execution governance structures tailored for both equities and fixed-income products. These included a best execution committee chaired by senior management and made up of representatives from the execution team, compliance function and other teams. Issues such as delays in execution, deviations in execution outcomes as compared against benchmarks and the execution failures of engaged third parties were discussed at committee meetings. The LC indicated that these discussions were helpful in its efforts to continuously improve its best execution arrangements. Best execution policies and procedures 3. Some LCs implemented comprehensive policies and procedures detailing the factors to be considered in delivering best execution, including applicability and carve outs, handling clients’ orders in the absence of pricing information or quotes and the responsibilities of execution staff and related control functions. 4. We observed the following practices which deviated from the expected standards.  In general, it appeared that LCs’ policies and procedures did not cover all types of products in sufficient detail. In particular, existing policies and procedures governing execution handling, monitoring and controls focused mainly on equities but not other asset classes such as fixed-income products.  Some LCs adopted relevant sections2 from the Code of Conduct as their best execution policy without further elaboration.  A number of the LCs had policies and procedures in place but could not demonstrate that they were subject to periodic review. Good practices The policies and procedures of one LC outlined the use of various benchmarks such as Volume Weighted Average Price (VWAP) and Implementation Shortfall to assess best execution. It clearly defined deviation thresholds which, when crossed, would warrant investigation and the escalation path for exceptions. Another LC’s policies and procedures referenced regional regulations and provided practical guidance to staff on how to adhere to these. The guidance defined best execution as well as controls to measure it both pre-trade and post-trade. It also outlined how to validate the quotes provided to clients in cases where multiple quotes for debt securities transactions could not be obtained by execution staff. 2 For example, one or more of paragraphs 3.1, 3.2 and 3.3 from the Code of Conduct were the only guidance specified in the policies and procedures.

5 Policies and procedures on disclosure of best execution arrangements 5. Some LCs made policies and procedures available on public websites to disclose their general execution arrangements. 6. One LC would provide the written policies and procedures on best execution to clients when they opened accounts as a way to disclose best execution arrangements including carve outs for specific products and trade scenarios, rebates and soft dollar arrangements, exclusive use of affiliates and preferred brokers. 7. Another LC required its staff to maintain order execution records showing the time of order receipt and execution, quotes obtained and other information to evidence best execution. These records would be provided to clients upon request. Good practices Some LCs had post-trade execution reports, eg, reports comparing the execution outcomes of their clients’ orders with benchmarks such as VWAP. Such post-trade execution reports were available to clients upon request. Training 8. Generally, LCs conducted best execution-related induction training for new staff, including staff from the control functions. Their training covered best execution policy, order handling procedures and conflicts of interest. 9. Some LCs provided periodic communications such as newsletters, periodicals and monthly meetings to provide staff with updates related to best execution. 10. We observed the following practices which deviated from the expected standards.  The majority of LCs only conducted training on an ad-hoc basis. The absence of regular training may fail to ensure staff awareness of best execution.  Some LCs did not maintain attendance logs or records to document the training conducted.

6 II. Best execution factors Expected standards Sufficient steps should be taken to obtain the best available terms when executing client orders, taking into account price, cost, speed of execution, likelihood of execution, speed of settlement, likelihood of settlement, size and nature of the order and any other relevant considerations. The relative importance of each best execution factor may vary from case to case and best execution of certain types of instructions should be assessed against multiple factors. Where a client has given specific instructions which cover one part or aspect of an order, this should not be treated as releasing LCs from their best execution obligations for other parts or aspects of the order. Observations 11. Some LCs listed in their policies and procedures different best execution factors, such as price, cost, speed and likelihood of execution, speed and likelihood of settlement and size and nature of order. 12. The LCs which specified execution factors all agreed that clients’ requirements were paramount and the relative importance of each best execution factor varied from case to case. For example, for an agency order for a small quantity of stock, price and speed of execution might be the most important factors, whereas for the execution of illiquid stocks, likelihood of execution might be more important. 13. We observed the following practices which deviated from the expected standards.  One LC, with institutional clients placing sizeable orders for illiquid stocks, adopted price as the only best execution factor, without considering other factors such as the likelihood of execution which might be more important.  Some LCs said they did not need to consider any best execution factors because they had simple business models which relayed client orders to the Stock Exchange for execution only. However, it was not uncommon to find instances where LC staff needed to exercise discretion in executing clients’ orders. For example, clients’ orders for illiquid stocks often needed to be split to reduce market impact, and the final execution price would be influenced by how the execution staff used their discretion to split and place these orders. Good practices An LC had policies outlining specific best execution factors relevant to its business. The policies prioritised factors across different clients (eg, corporate clients or individual clients), asset classes (eg, equities or debt securities) and trading scenarios (eg, in the absence of liquidity or multiple quotes).

7 III. Applicability of best execution Expected standards When LCs enter into agency or back-to-back principal transactions3 with clients, the obligation to deliver best execution remains with LCs where clients rely on LCs to protect their interests in order execution. When LCs enter into principal transactions with clients, other than those which are back-to-back in nature, LCs should carry out their own assessments to determine whether clients are relying on LCs to protect their interests, and whether the best execution obligation is owed, before applying any carve outs. Observations 14. In general, LCs considered best execution to be applicable to all agency and back-to-back principal transactions when clients rely on LCs to protect their interests. For example, best execution applies when an LC sources quotes on a debt security from different counterparties on behalf of its client. 15. LCs had different ways to assess whether they owed best execution duties in principal (proprietary) transactions where they were market makers providing quotes to clients. Some LCs systematically assessed whether best execution applied as illustrated in the Good practices below. 16. We observed the following practice which deviated from the expected standards.  Some LCs inappropriately carved out all OTC products including debt securities without assessment on the basis that market data for these products was not easily accessible.  Without assessing reliance from clients to act in their best interests during the execution process, some LCs considered that they did not owe best execution duties to clients in principal (proprietary) transactions. 3 Under paragraph 8.3 part A of the Code of Conduct, back-to-back transactions refer to those transactions where a licensed or registered person, after receiving – (a) a purchase order from an investor, purchases an investment product from a third party and then sells the same investment product to the investor; or (b) a sell order from an investor, purchases an investment product from the investor and then sells the same investment product to a third party, and no market risk is taken by the licensed or registered person.

8 Good practices Some LCs used a more systematic way to assess whether best execution applied. The following flow chart illustrates how they made this assessment. Assessment criteria (to be considered collectively): • Does client initiate the transaction? • Is it a market convention for clients to “shop around”? • Is it a relatively transparent market? • Is disclosure made to client that no best execution is provided? Best execution not applicable Best execution applicable Limited information Yes Relationship with client: Agency / principal (back-to-back) Relationship with client: Principal (proprietary) No

9 IV. Responsibilities of execution staff Expected standards Execution staff and their supervisors form an important first line of defence in the trade execution process. Reasonable diligence should be demonstrated in handling client instructions, monitoring execution outcomes and, where applicable, taking steps to obtain multiple quotes. In the absence of multiple quotes, execution staff should obtain sufficient pricing information on a best-effort basis to validate quotes provided to clients. Observations Handling client instructions 18. Cases where execution staff might occasionally exercise discretion in executing client orders were common. For control and monitoring purposes, one LC required execution staff to properly record all orders requiring their discretion. 19. We observed the following practice which deviated from the expected standards.  For orders to be split for execution, the execution staff of one LC would record the original order quantity on paper upon receiving a client’s instructions and then manually cross out the quantity executed. Such procedures are prone to human error. We noted instances where the execution staff did not properly record orders which had not been executed and as a result failed to execute them. Responsibilities of execution team supervisors 20. The supervisors of one LC’s execution team adopted a systematic approach to monitor order execution, reviewing it against client instructions and benchmarks specific to best execution at the end of each trading day. 21. We observed the following practice which deviated from the expected standards.  An LC with high daily execution turnover indicated that senior members of the execution team did not take a systematic approach to ensure execution quality and only performed an “eye-ball” and ad-hoc check on all executed orders. Obtaining multiple quotes 22. Some LCs had policies and procedures in place to specify the minimum number of quotes to be obtained by execution staff for debt securities and structured products. Some LCs could provide evidence to demonstrate that the prevailing best bid and offer prices had been obtained. Good practices One LC would obtain quotes from different counterparties for debt securities transactions. In the trading system, the execution staff were required to input the traded price and the next best available quotes obtained. Apart from checking the completeness and accuracy of the quotes inputted into the trading system, compliance staff would also perform post-trade checking to compare quotes and execution prices and determine if orders were executed at the prevailing best bid and offer prices.

10 V. Controls and monitoring Expected standards Controls and monitoring carried out by second and third lines of defence, such as compliance and internal audit functions, should be in place to review the quality of execution and to detect and address anomalies. Based on the characteristics of different financial instruments and the complexity and scale of the LCs operations, it should determine appropriate metrics and reference benchmarks to assess execution quality. Observations Second and third lines of defence 23. We observed the following practices which deviated from the expected standards.  The tools used by the compliance and internal audit functions of a number of LCs to assess execution quality for OTC products were generally less developed than those for exchange-traded products.  The compliance functions of a number of LCs did not appear to have defined roles for the control and monitoring of best execution. In-depth and rigorous independent challenge by compliance functions was insufficient throughout the order execution process.  Although a number of orders required that execution staff exercise discretion to complete them over a period of time during trading hours, the controls and monitoring implemented by an LC only assessed the timeliness of order completion in determining best execution without considering price and other factors which might be more important.  One LC highlighted that the duty of ensuring best execution did not extend beyond the supervisors of the execution team. This LC’s compliance function considered that its responsibilities in this area should be solely focused on operational errors and market misconduct.  The internal audit functions of some LCs did not conduct reviews relating to best execution. Good practices An LC’s compliance function reviewed execution reports daily. The reports compared all execution prices with VWAP. Execution staff were required to explain deviations above certain thresholds and exceptions would be escalated to senior management for review. Another LC’s compliance function checked samples on a daily basis. Compliance staff would compare the time the orders were placed with the time they were executed and listen to the audio recordings of clients’ instructions to determine whether the orders were executed per clients’ instructions and identify any delays. Another LC’s internal audit function periodically reviewed best execution policies and procedures and assessed the effectiveness of the control framework. The periodic review assessed the appropriateness of the benchmarks used by execution staff to measure execution outcomes and reviewed trade errors and remediation measures.

11 Monitoring trading tools provided to clients 24. In compliance with the Code of Conduct4 , the algorithms of LCs’ trading tools were regularly and robustly tested to ensure they operated as intended. Some LCs provided training and support to clients to ensure they understood the use and objectives of the algorithms. 25. Some LCs had Smart Order Routing in place which helped achieve better execution by allowing access to several venues and automated searches for better execution prices. 26. In determining whether best execution had been achieved, one LC conducted daily reviews to compare the algorithm’s performance against benchmarks and also periodically collected client feedback on the algorithm’s effectiveness and efficiency. 27. We observed the following practice which deviated from the expected standards.  One LC indicated that it did not owe any best execution obligations to clients who placed orders via the trading tools it provided, as the orders were placed without the LC’s intervention. This LC did not monitor the execution outcomes of orders placed via its trading tools. Transaction cost analysis 28. Some LCs used transaction cost analysis (TCA) reports to monitor execution quality by comparing execution outcomes against benchmarks such as VWAP or Implementation Shortfall. Good practices A range of bespoke TCA reports were calibrated by various LCs based on their clients’ trading strategies and requirements. A number of LCs with more sophisticated international operations made use of TCA reports, benchmarks and other systematic and automated tools to monitor execution. Some LCs used TCA reports to detect abnormal trends and fine-tune their execution strategy to achieve better results for clients. 4 Paragraph 18.10 and Schedule 7 of the Code of Conduct.

12 VI. Arrangements with affiliates, connected parties and third parties Expected standards Due diligence should be carried out on affiliates, connected parties or third parties engaged for execution, and a systematic process should be in place to continuously monitor execution outcomes. LCs should take action to ensure that execution arrangements with affiliates, connected parties and third parties do not undermine the delivery of best execution. Regardless of whether orders are executed through affiliates, connected parties or third parties, the obligation to deliver best execution remains with LCs. Observations Selection and monitoring 29. A number of LCs engaged affiliates, connected parties or third parties to execute orders for their clients for various reasons, such as to execute orders for OTC products or overseas equities which were not directly accessible by the LCs. In general, they were able to demonstrate adequate selection criteria. 30. LCs’ monitoring and controls for the selection of third parties differed. They also had different responses where the execution quality of third parties did not meet their expectations. In the case of a third party’s poor performance, such as frequent execution failure, one LC would ban it for a period of time, reduce future transaction volume directed to it or even remove it from the approval list for execution. 31. We observed the following practices which deviated from the expected standards.  A number of LCs did not conduct ongoing monitoring or assessment of the execution quality of the affiliates, connected parties and third parties engaged for execution.  Some LCs did not have controls and monitoring for orders placed with affiliates and connected parties for execution. For example, one LC was unaware that its orders placed with affiliates for execution might be executed in an alternative liquidity pool. Good practices Some LCs had robust and systematic processes in place to review third party execution. Apart from reviewing TCA reports, execution staff would also evaluate the performance of third parties by assigning them scores, thereby creating a feedback loop to promote execution quality. An LC had mechanisms in place to compare actual transaction volume with third parties against their rating as assessed by the execution team. It also established a committee which was given decision-making responsibility and authority in relation to the global selection, appointment and ongoing risk monitoring of all third parties appointed for execution.

13 Inducement arrangements 32. A number of LCs understood the implications of inducement arrangements with affiliates, connected parties or third parties. They stressed that obligations to deliver best execution to their clients remained the key priority. Soft dollars, rebates and other inducement arrangements would be clearly disclosed to clients in accordance with applicable regulatory requirements. 33. We observed the following practice which deviated from the expected standards.  Without other further assessment, one LC indicated that anticipating reciprocal order flows was a key consideration when selecting third parties for order execution.