1. Introduction

This Staff Notice is a follow-up to Joint CSA/IIROC Consultation Paper 23-406 Internalization within the Canadian Equity Market (the Consultation Paper) that was published for a 60-day comment period on March 12, 2019, by staff of the Canadian Securities Administrators (CSA) and the Investment Industry Regulatory Organization of Canada (IIROC) (together, Staff or we).1The Consultation Paper was published in response to concerns raised about the internalization of equity trades on Canadian marketplaces. 21 comment letters were received.

This Staff Notice summarizes the feedback received, refreshes certain data published as part of the Consultation Paper and provides an update on next steps.

1. Background

Beginning in 2017, Staff became aware of growing concerns about a perceived increase in the magnitude of internalization of retail/small orders within the Canadian equity market.

Internalization generally refers to trades that are executed with the same dealer as both the buyer and the seller, with the dealer either acting as an agent for its clients on both sides of the trade, or trading as principal and taking the other side of a client order. Internalized trades occur on Canadian marketplaces as either “intentional” or “unintentional” crosses.2

The Consultation Paper provided background information that described certain relevant aspects of the Canadian rule framework, identified specific issues and concerns, and provided data illustrating recent levels of internalization in Canada.

Issues and Concerns

Below, we discuss the primary issues presented and the feedback received in response to the Consultation Paper. A complete summary of comments received and Staff responses is at Appendix B.

1. Broker Preferencing
   As described in the Consultation Paper, broker preferencing is an important element of the concerns raised in relation to internalization. Broker preferencing is a common order matching feature of many Canadian equity marketplaces. It allows an incoming order sent to a marketplace to match and trade first with other orders from the same dealer, ahead of orders from other dealers that are at the same price and which have time priority. Broker preferencing is relevant to issues associated with internalization as it can facilitate internalization through the execution of unintentional crosses. It has been a divisive issue for many years in Canada, and the responses that Staff received to specific questions in the Consultation Paper related to broker preferencing reflects the continuing divergence in the views of stakeholders.
   
   Some respondents articulated their belief that broker preferencing is a benefit to clients of dealers and a preferable alternative to equity market structure models in other jurisdictions. Some supporters expressed the view that retail clients were specific beneficiaries of better execution quality as a result of broker preferencing, and that the ability for dealers to efficiently interact with their own orders on a marketplace encourages the transparent display of liquidity on Canadian marketplaces.
   
   Other commenters however, described negative impacts of broker preferencing, notably in the context of fairness through the creation of an unlevel playing field, where not all market participants, including investors, have equal access to interact with orders. Despite the views that broker preferencing benefits the Canadian market by encouraging displayed liquidity, some respondents argued that the impact is less beneficial and felt that the ability to override the time priority of other displayed orders in an order book results in a negative impact on immediacy and a perception of a lack of fairness where a displayed order might not receive an execution despite it having been at the top of the order book queue.
   
   The Consultation Paper specifically requested views on whether broker preferencing conveys greater benefits to larger dealers. Most commenters agreed that larger dealers and their clients may receive greater benefits. The Consultation Paper also specifically requested any data that illustrated either the positive or negative impacts of broker preferencing (and internalization, more generally). Very limited data was received that could quantitatively evidence the impacts.
2. The Individual Versus the Common Good
   The Consultation Paper described the issue of the individual good versus the common good. It was noted that, while it may be reasonable to conclude that the internalization of client orders may benefit individual dealers and their respective clients, it may also be true that a market in which participants collectively act to maximize their own benefits may not result in a market which functions in the best interests of all those participating. Staff noted the importance of a balance between a market that adheres to the principles of fairness and integrity and one that operates to the benefit of the individual participants who interact within it.
   
   The comments received regarding the common versus the individual good were mixed. Many characterized internalization as being contrary to the common good, while others suggested that Canadian market structure should seek to find an appropriate balance through the use of internalization.
3. Segmentation of Retail Orders
   Segmentation of orders typically means the separation of orders from one class or type of market participant from those of other classes of participants. In describing this issue in the Consultation Paper, Staff noted that, in the Canadian context, this is typically focused on the orders of retail investors. The Consultation Paper discussed the value proposition inherent in interacting with retail orders, and we offered commentary on how Canadian market structure has evolved with various methods that seek to either implicitly or explicitly segment retail orders.
   
   Most commenters believed that the segmentation of orders is a concern for a variety of reasons, including that the removal of access to retail orders (or orders of any participant) is contrary to principles of fairness and may result in a lower quality, less liquid and less competitive market. Some felt that a distinction was warranted between the segmentation resulting from participants choosing between various commercial models that are available to all market participants, and the segmentation schemes that serve to isolate retail orders through restricting access.
   
   It was not evident from responses to the Consultation Paper that the Canadian market has reached a point where the level of segmentation requires an immediate policy response. Most respondents believed that the structure of the Canadian market provides for favourable outcomes for retail investors, although continued caution was recommended to avoid unbalanced results.
4. Automated Matching Against Client Orders on a Marketplace
   The Consultation Paper highlighted that, as part of the ongoing technological evolution of the Canadian market, systems may be used by dealers to automate the internalization of orders through broker preferencing. It was noted by Staff, that such systems may appear to exhibit the characteristics of a marketplace as defined within the Canadian rule framework.3
   
   Most commenters were of the view that systems that automate the internalization of orders should be considered a marketplace, and that relevant provisions of the rules should apply. Concerns were raised about the creation of discrete silos of liquidity within dealers that become inaccessible to the broader market. Some, however, suggested that dealers may simply be automating what has historically been a manual process, one that has never been considered a marketplace, and that the application of technology alone should not change the regulatory classification of dealer workflows.

1. Revised Internalization Data

In addition to describing various issues and seeking feedback, the Consultation Paper also included data that explored the magnitude of:

• intentional crosses;
• unintentional crosses;
• crosses where the dealer acted as principal; and
• the use of broker preferencing on certain Canadian marketplaces.

With respect to intentional and unintentional crosses, the data in the Consultation Paper relied on information received by IIROC through the Market Regulation Feed and submitted by each Canadian marketplace for the period of January 2016 through June 2018.

Data examining the magnitude of broker preferencing was provided directly to Staff by the marketplaces themselves. However, not all Canadian marketplaces were able to accurately distinguish between unintentional crosses resulting from broker preferencing itself (and where time priority was not followed), and unintentional crosses where a resting order was already in a position of time priority and would have been executed despite the availability of broker preferencing. This incomplete broker preferencing data requested from marketplaces covered the period of January 2017 through July 2018.

Some time has passed since the publication of the Consultation Paper and Staff’s review of the feedback received and associated data collected, and we are of the view that it is important to update certain data to more accurately reflect current market statistics. Therefore, we are republishing certain data at Appendix A that updates the period of coverage from January 2016 through October 2019. We have also added charts that represent information that the Consultation Paper included in graphs to make the information easier to read. 

The data at Appendix A however, does not update the specific broker preferencing information initially provided by marketplaces for purposes of the Consultation Paper. While this data was informative, it did not include all Canadian marketplaces and as such, is incomplete for the purposes of regulatory policy decisions. IIROC has been working with Canadian marketplaces to receive broker preferencing data as part of the Market Regulation Feed, but IIROC has not received this for a sufficient length of time to provide updated information at Appendix A. Future analysis will consider this information and may also consider other market structure developments such as changes implemented by marketplaces that may impact levels of internalization.

1. Next Steps

The Consultation Paper purposely did not offer Staff’s views on the issues presented, but rather, focused specifically on seeking feedback in order to help inform future policy decisions. The background information and related narrative in the Consultation Paper recognized the competing interests associated with internalization and attempted to provide a balanced presentation of what Staff considers to be the primary issues.

The feedback received was varied and, consistent with the way the issues were framed in the Consultation Paper, balance was a common theme presented in the responses. Specifically regarding broker preferencing, while the practice is at odds with price/time priority in order execution, broker preferencing is a longstanding part of Canadian market structure. As currently functioning, broker preferencing may allow dealers to benefit from interaction with their own orders, and may also benefit individual clients with improved execution quality. There may be nuanced outcomes of broker preferencing, and some market participants may not be impacted in the same way as others. Based on the feedback received and the data reviewed, we do not believe that the Canadian market is presently functioning in a way that warrants near-term policy work or changes to the current rule framework.

As noted, the Consultation Paper highlighted that systems may be used by dealers to automate the internalization of orders, and that these systems may appear to exhibit characteristics of a marketplace as defined within the Canadian regulatory framework. This is further described in the guidance included in the Companion Policy to National Instrument 21-101 Marketplace Operation (NI 21-101CP) regarding when dealers may be operating a marketplace.4 The CSA will consider whether additional clarification should be provided in relation to when a system is a “marketplace”.

With respect to the updated data published at Appendix A, Staff note that the level of unintentional crosses has increased since the six-month period of January through June 2018, which was the final period of data initially published alongside the Consultation Paper. While the most recent data illustrates an increase, Staff have looked at the underlying non-public data and are comfortable that the increase is not an indication of broad changes in the way in which dealers are managing their orders or of a specific concern that necessitates an immediate regulatory policy response.

We will however, continue to monitor the data on an ongoing basis and if there are any indications that changes to internalization practices, including internalization that is enabled through the use of dealer systems, are possibly impacting Canadian market quality in a negative way, we will consider appropriate responses at that time.

1. Questions

Please refer your questions to any of the following:

Appendix A – Quantitative Analysis of Internalization on Canadian Marketplaces

Appendix B – Summary of comments received and responses

• 1Published at: https://www.osc.gov.on.ca/documents/en/Securities-Category2/csa_20190312_internalization-within-the-canadian-equity-market.pdf
• 2An “intentional” cross is considered to mean a trade that results from the simultaneous entry by a dealer of both the buy and the sell sides of a transaction in the same security at the same price. An “unintentional” cross is considered to mean the execution of a trade where the two orders are from the same dealer, but not simultaneously entered.
• 3The definition of a “marketplace” is included in National Instrument 21-101 Marketplace Operation and, in Ontario, also in the Securities Act (Ontario).
• 4Specifically, subsection 2.1(8) of NI 21-101CP clarifies that, if a dealer uses a system to match buy and sell orders or pair orders with contra-side orders outside of a marketplace and routes the matched or paired orders to a marketplace as a cross, it may be considered to be operating a marketplace under subparagraph (a)(iii) of the definition of “marketplace”.