In the popular movie series “Back to the Future” the villain is a school bully named Biff Tannen. At the end of the first movie, he gets punched in the face by Marty McFlys father to enable the timeline to be restored. In the second movie, Biff goes back in time, and, with a sports almanac, uses it to bet his way to wealth and power. In the end, he is stopped by Marty, but not before sobering views of a future changed by Biffs actions.
In the ongoing saga of market structure debates over speed-bumps, I see a parallel between the CHXs attempt to introduce a speed-bump (to provide advantages to its market makers) and Biff. Their first proposal, for an asymmetric delay, was met by the equivalent of a punch in the face by commenters who argued that it would distort the markets. Unfortunately, their second proposal, which would provide a symmetrical delay, but one that excludes their best market makers, is arguably worse. This new proposal is a step backwards in time and is reminiscent of the structural advantages held by specialists on the old NYSE[1] or Nasdaq market makers on the old Selectnet system[2].
On the surface, the CHX proposal makes sense. It could turn out that incentivizing “obligated” market makers to quote more will improve liquidity and price discovery, but, then again, it might not. From the CHX perspective, becoming the first modern exchange to re-introduce structural advantages for market makers, could boost their market share and attract incremental activity. Unfortunately, due to a couple of outdated regulations (Rule 605 and the Order Protection Rule in Reg NMS), this proposal would be quite problematic. In addition, if it succeeds, we should expect the other exchanges to introduce similar models, magnifying the extent of the problems it could create.
The specific CHX proposal is to provide lead market makers (who have obligations to maintain defined statistical levels of two way quotes in all securities they are designated in) the ability to place liquidity providing orders as well as to cancel orders without restriction, while the CHX will uniformly apply a 350-micro-second delay to all other order placement and cancellation actions on the exchange.
This provides lead market makers an order placement advantage which would let them legally front-run other orders to provide liquidity at the front of order queues. In practice, whenever a lead market maker sees a price change on the other exchanges, they will be able to react and place re-priced orders on the CHX to gain primary position in the order queue while other participants (including client algorithms) are delayed. The ability to be the first quote, translates directly into more profit for the market maker, while the orders that were held back will suffer adverse selection. While this may end up as a worthwhile market innovation if it improves liquidity, it would be terrible policy for all exchanges to facilitate this sort of two-tiered market.
The proposal also confers a speed advantage for order cancellation and that is more problematic, under the current regulatory regime. The CHX proposal would allow their qualifying market makers to move their quotes AFTER an inbound order was sent to the exchange matching engine, whenever they believed that such a trade could lose them money. In practice, this means that a CHX quote, if it was the last quote among the 13 exchanges, would likely be impossible to access as the market maker would have strong incentive to move it. There are two problems with this. First, due to the Order Protection Rule, routing firms are required to send orders to access these quotes. Thus, an order routing firm, despite compelling evidence that in certain situations they will not receive an execution, will need to route to the CHX anyway. To make matters worse for routing firms, in addition to the time spent routing to the CHX, orders sent there will be delayed, translating into an increased risk of missing out on liquidity elsewhere. It is worth noting, however, that if there was no Order Protection Rule, routing firms would be able to decide if the potentially increased liquidity provided by the CHX market makers was worth the delay. This would then mirror the value proposition of the NEO exchange model in Canada, since that exchanges quotes are not protected.
Second, the ability to cancel displayed quotes during the time inbound orders are delayed distorts the NBBO provided by the SIP, which is designed to only include executable quotes. As I explained, the lead market makers would have a strong incentive to cancel their quotes whenever they see that they are the only quote or are about to become the only quote among all the exchanges at a price level. This would, in turn, provide the appearance of fake liquidity which will make the SIP quote less reliable. This is a problem because, under rule 605, all market centers report best execution statistics based on the SIP quote. Retail investors, in particular, would be harmed by this as their marketable orders are sent to wholesale market makers that are judged against the SIP quote. If the quote became less reliable, the amount of price improvement received by retail investors would be jeopardized.
It is important to note that the ability of market makers to move displayed quotes during the speedbump is in direct contrast to IEX, which does not allow any repricing or cancelation of displayed quotes during their delay. In fact, this point was made by several commenters, including myself, when discussing the SECs proposal to interpret Automated Quotations Under Regulation NMS. As I said in that comment letter: Exchange re-pricing logic within a delay would render such orders conditional upon the market data feeds that change during the delay. This would result in precisely the kind of “maybe” quotations Rule 611 was designed to prevent.[3]
To be clear, the issue with the CHX proposal is not that a market maker centric system is a bad idea, but rather that allowing such an exchange to have their quotes protected by the government is wrong. If the OPR were abolished, and best execution rules were based on a firm National Best Bid and Offer, which excludes quotes (like these) that are not fully executable, I would have no objection to CHX trying this as an innovation. Unfortunately, both rules exist, and lead to the inescapable conclusion that the CHX proposal should be rejected.
Hopefully, the SEC will stop this attempt, or we might need a time machine and a real-life Marty McFly to go back and undo the damage…
[1] Specialists had exclusive access to the full order book and matched all orders themselves at their post
[2] Nasdaq market makers had the ability to react to institutional trade requests with decline and move; The only automated execution was via SOES (Small Order Entry System) and the market makers famously referred to retail sized traders who executed orders against them at prices they thought were stale as SOES Bandits
[3] https://www.sec.gov/comments/s7-03-16/s70316-18.pdf
David Weisberger isPresident of Exquam LLC
