Notice of Approval – Lynx ATS – Periodic Matching – Tradelogiq Markets Inc.

COMMENTS

TRADELOGIQ RESPONSE

Both commenters indicated general support for the model, subject to certain caveats and/or suggestions. The approach being taken by Lynx to leverage existing connectivity was also acknowledged as helpful, as it reduces overall disruption for dealers.

One of the two commenters also indicated support for allowing marketplaces to innovate and compete for order flow, and if necessary, even fail.

We agree. Innovations such as those that comprise the Lynx periodic matching model should be allowed to proceed so long as they do not violate regulatory requirements or principles. Because Lynx is currently and will remain unprotected, subscribers will continue to not be subject to any obligations under order protection requirements to route their marketable orders to trade against displayed quotes on Lynx. In addition, subscribers can choose if and when to interact with the model based on whether it is beneficial to do so for the execution of their and/or their clients’ orders.

Both commenters suggested revisiting the EOC Final Turn on the basis that it would result in EOCs in that second phase of matching receiving a better price than EOCs participating in the first EOC-to-DAY phase of matching, the latter potentially having paid the spread.

Suggested alternatives included allowing all orders to trade at the midpoint, matching EOC orders with other EOC orders first before allowing EOC-to-DAY interaction, or to cancel any unfilled EOC orders after the EOC-to-DAY matching phase.

We acknowledge that there may be circumstances where a later-arriving EOC receives an execution at a price that is better than the earlier-arriving EOC, but do not view this as unreasonable or unfair when considering the intent of the EOC Final Turn (as was outlined in the Notice), and what we believe will result in a positive overall effect.

We had considered alternatives such as those suggested by commenters. However, in our view, the alternatives either unfairly penalize DAY orders that have contributed to the displayed volume on Lynx or under-incent provision of liquidity by making it less or even not worthwhile to post unless willing to provide at least midpoint price improvement.

As indicated in the Notice, one of the purposes of the EOC Final Turn is to help to mitigate the risk that DAY volume is no longer available at the time of the Match Event, whether because it faded before, or was exhausted during, the Match Event. This type of situation is one of the common concerns associated with market models that incorporate delays.

We also do not believe it is an unfair or unreasonable outcome considering the earlier-arriving EOC has the benefit of higher fill certainty than a later-arriving EOC.

Further, it is our expectation that the experience for any given sender of an EOC may differ from EOC to EOC and from Match Event to Match Event, with the benefits of being earlier-arriving or later-arriving balancing out over time, leading to a better overall experience.

Lastly, we feel that the innovation of the EOC Final Turn will result in access to a novel source of liquidity and increased price improvement for active order flow that would not otherwise be available if unfilled EOC orders were simply cancelled back.

Both commenters raised questions about the proposed display price mechanism in the event that it results in orders being displayed at prices that do not reflect their more aggressive tradeable price.

One commenter suggested that this is misleading, but this commenter also noted that as an unprotected market, the pricing on Lynx would not contribute to the protected NBBO and will not impact the accessibility or price discovery of protected quotations on other markets.

The other commenter expressed the view that displayed quotations on all marketplaces are expected to represent a source of price discovery and are inputs into others’ activity on all markets, and therefore was of the view that displaying orders at a price that is less aggressive than their tradeable price would be misleading to participants who may be looking at Lynx as a source of price discovery.

We appreciate commenters’ concerns about the potential for the display price mechanism to result in misleading prices, and note that the two commenters appear to have differing views as to the significance of this facet of the model based on their views on the extent to which visible pricing on Lynx contributes to price discovery more broadly.

It is not our expectation that subscribers will frequently enter visible DAY orders at prices that are at or through the Protected NBBO midpoint. As a result, we expect it to be the exception rather than the norm that booked DAY orders in the Visible Book will be displayed at prices that do not reflect their more aggressive executable prices, and we believe that this outcome is reasonable when considering that the purpose of the mechanism is to manage the potential for the appearance of locked or crossed DAY orders received during a Collection Period.

Further, we note that the outcome in question, whereby the volume of certain price-improving orders will be displayed at a less aggressive price, can currently also arise on NEO-N, because the volume of more aggressively priced buy (sell) midpoint peg volume on NEO-N is displayed as part of the total aggregated volume at the less aggressive best NEO-N bid (ask).

On both NEO-N and Lynx, there can therefore be circumstances where the volume displayed at the venue’s best bid/ask includes the volume of orders that have a more aggressive tradeable price. Proper disclosure by the venue should allow subscribers to understand that the volume displayed at the best bid/ask on NEO-N, or displayed at or just inside of the Protected NBBO midpoint on Lynx, represents volume available at that displayed price or better.

We also note that OSC staff have imposed terms and conditions relating to the concerns raised, as set out above.

Both commenters suggested the implementation of a freeze period at the end of the collection phase, with one suggesting this could be combined with a randomization of the collection periods. It was suggested that this could help to reduce the potential for gaming and quote fading in relation to the displayed orders, and resulting in better execution quality for both EOC and DAY orders.

We appreciate the commenters’ suggestion regarding the inclusion of a freeze period, and agree that it could be beneficial. However, we prefer to proceed with the launch as proposed in order to allow us to assess whether the model is delivering the intended outcomes, and will consider adding a freeze period depending on user experience and the feedback received post-launch.

Clarification was requested as to whether the period between Match Events would be implemented with a static or random interval, indicating a preference for a random interval.

We are currently intending to implement with a randomized period between Match Events – which will be effected through the combination of a static duration and a randomization window as indicated in the Notice. For example, if the static duration is set at 4ms with a randomization window of 1ms on each side, Match Events will occur on a randomized basis every 3ms to 5ms. Notice regarding the durations to be applied will be provided at least 2 months prior to launch.