Will the repeal of Reg NMS’s Order Protection Rule put the genie back in the bottle?
By Hitesh Mittal, Founder & CEO, BestEx Research
The SEC has proposed repealing the Order Protection Rule and its related rules on locked and crossed market prohibitions1 (together referred to as “OPR” in the rest of the note), two of the foundational components of Regulation NMS. There are arguments for and against the OPR. The strongest argument in favor of repeal is the unnecessary fragmentation it may have created, allowing new exchanges to emerge because, as long as they can attract market makers and establish the best quote, liquidity takers cannot ignore them. The strongest argument against repeal is the potential for locked markets and a less meaningful NBBO.
Both sides have merit in their arguments. But I think the framing of the debate is wrong, and the question worth asking is a different one. What will unwinding OPR actually do to US equity markets?
My answer, in short, is that a repeal of OPR will change the terms of competition among venues far more than it changes the number of them. OPR was the direct reason for one wave of venue creation, the maximize-the-rebate exchanges. It was only the indirect catalyst for the creation of the rest of the venues—the ATSs, SDPs, and some of the exchanges. Venues built around internalization, adverse selection, access-fee avoidance, listing law, trading hours, and whatever tokenization turns into will still have their reason to exist the day after repeal. The genie is not going back in the bottle; we will still have fragmentation. It will only evolve into something else.
That claim only makes sense considered against history, so it is worth reviewing how we got from two exchanges to sixteen, 30 ATSs, and numerous SDPs in the first place.
How OPR let the genie out
The end of duopoly.
With the introduction of OPR, we first witnessed the breaking of the NYSE and Nasdaq duopoly. ARCA, BATS, and Direct Edge all directly challenged the incumbents. This was largely expected and perhaps the intended effect of OPR.
Traditionally, the primary challenge in starting an exchange was attracting order flow, and market makers will show up as long as you have it. OPR changed the game by requiring trading centers to prevent executions at prices inferior to the NBBO. As long as a liquidity provider is willing to quote at an exchange, the takers cannot ignore them. The game therefore became increasingly about attracting liquidity providers.
The maximize-the-rebate strategy (and access fee).
As a result, these new exchanges relied heavily on a maximize-the-rebate strategy. In a competitive environment exchanges generally make 2-5 mills per share net on two sides of the trade. They can do so by charging both the suppliers and takers a fee, or they can do so by offering a rebate to one side and a higher fee to the other side to compensate for the loss and include their margins. To determine who to favor, exchanges study the elasticity of liquidity suppliers and takers—how much more volume can they expect from that side if the pricing is reduced. Since the elasticity of liquidity takers is low (due to the OPR), most exchanges charged the highest fee possible to liquidity takers, and since the elasticity of liquidity providers was very high (they could provide liquidity on any number of exchanges), exchanges offered the highest rebates for providing liquidity.
The playbook, then, was simple. Attract the largest market makers, HFT firms, and broker-dealers to quote on your exchange by offering the highest rebates possible (or even better, an equity stake). Once those firms established competitive protected quotes, OPR ensured that liquidity takers could not ignore them.
But it is important to note what this playbook does not require: order flow, a differentiated matching engine, or a single institutional client relationship. It requires a pricing structure with a high rebate, relationships with the largest market makers and a rule that makes their quotes unavoidable.
An unleveled playing field.
Rebates, however, did not solve the adverse selection problem, which HFT market making firms care about equally. Exchanges “innovated” by creating increasing levels of asymmetry between their largest customers, the HFT firms, and everyone else. They offered volume tiers, proprietary data feeds, increasingly sophisticated order types, and faster access. HFT firms also invested heavily in low-latency technology to detect when the market was about to move and cancel their quotes before potentially price-moving orders could trade against them. While the same advanced functionality was offered to everyone, only the largest HFT firms with tens of millions of dollars budgeted annually could make the most optimal use of it.
Low-price liquid stocks create long queues and invite inverted exchanges.
Another distortion created by OPR was artificially large spreads for low-price, highly liquid stocks. OPR required2 a harmonized tick size across exchanges, which was set to one penny for all stocks trading above $1. For example, large-cap stocks trading around $5 traded at 20 basis points of spread while similar large caps at higher prices traded at roughly 3 basis points. Access fees3 alone added another 12 basis points of effective cost on those low-price names.
Queues at the best price became very long, so a liquidity provider joining the back of the queue had little realistic chance of trading. Inverted exchanges were born to solve this problem. By paying the liquidity taker rather than charging it and by charging the suppliers they reduced the queues giving providers a chance to fill quickly and lowered effective spreads for the taker. Note that while OPR requires harmonization of tick size, there is no proposal for the marketplace to freely determine its own tick size and tick sizes are slated to reduce to half a penny for liquid stocks by November 2027.
High access fees chart the course for ATSs and SDPs.
High maximum access fees, coupled with OPR, motivated institutional broker-dealers to innovate their business models. Institutional broker-dealers invested heavily in building their own ATSs with lowering of routing costs as one of the major motivations. Most large broker-dealers modified their execution algorithms to take within their own ATSs first. They simultaneously solicited HFT firms to provide liquidity within those ATSs. While only a few ATSs existed prior to Reg NMS, and were largely buy-side-to-buy-side crossing networks facilitating large transactions at the midpoint, most ATSs post-Reg NMS became internalization mechanisms for broker-dealers with average trade sizes similar to exchanges. Many of these ATSs today have even larger market share than exchanges.
With institutional broker-dealers becoming extremely cost conscious, HFT market making firms also sensed the opportunity and started offering their liquidity for "free" in the form of unregulated venues called single-dealer platforms (SDPs). Broker-dealers simultaneously invested in building smart order routers (SORs) to lower their routing fees. Most SORs were configured to take first in their own ATS, then inverted exchanges, then SDPs, then other ATSs, and only then regular maker-taker exchanges. And that routing hierarchy became the norm.
Retail flow leaves exchanges entirely.
While wholesalers existed before the Reg NMS era, the race for retail flow became even more heated post-Reg NMS with technological advancements, increased fragmentation and increased synergies among various kinds of flows these wholesalers operated. Retail brokers earned rebates on both limit orders (at exchanges) and market orders (in the form of payment for order flow) and zero commission trading was born. Exchanges became the last stop for both institutional and retail liquidity-taking order flow.
Addressing adverse selection with innovation.
While exchanges became the liquidity of last resort, OPR ensured that their quotes could not be ignored, so the distinct advantage they held over ATSs and SDPs was that exchanges were the only type of venue where prices could not be traded through. On the other hand, the excessive number of venues and being last place in the routing table meant that liquidity provision on exchange came with high adverse selection, for both liquidity providers and institutional broker-dealers.
IEX was one of the first to innovate in the exchange space to address the adverse selection issue, roughly a decade after Reg NMS arrived. Its best-known innovation is a 350-microsecond speed bump. HFT firms had spent enormous amounts of money building technology that allowed them to identify when a quote was becoming stale and cancel before an incoming order could pick them off. IEX, with its delay and integrated signal, was able to lower adverse selection for liquidity providers. Other exchanges, for example Nasdaq with MELO, have since created similar delay mechanisms for the same purpose.
More recently, ATSs have also started segmenting liquidity by grading liquidity taker flow to allow liquidity providers to optimize the tradeoff between fill rate and adverse selection. Trajectory crosses have also become increasingly popular as they cross orders at interval VWAP prices, thus minimizing adverse selection. Pre-Reg NMS there were three dark pools operating as ATSs, and today there are roughly thirty, each with customizable segments of liquidity and at least eight running their own trajectory crosses. New exchanges pursuing a maximize-rebate strategy also continue to appear, with MEMX and MIAX as the most recent examples.
This is the section of the history that matters most for the forecast. While OPR may have started it, many of the later innovations were introduced to address adverse selection, and they don’t necessarily depend on OPR to draw order flow anymore. Innovations such as speed bumps, segmented pools, and trajectory crosses are all just as valuable the day after a repeal as the day before.
What repeal actually changes
In my opinion, the biggest direct change we can expect if OPR is repealed is in the elasticity of liquidity takers and liquidity providers, which will lead to different incentive structures. Liquidity takers without obligation to route to the best-priced exchange (but with a very high access fee) will now have the ability to lock a quote.
With no obligation to take liquidity at an exchange even when it is quoting the NBBO, exchanges will have to work harder to win order flow. If there are two exchanges, one with a heavy maker-taker model and the other with a low-fee model, a liquidity taker can simply provide at the low-fee venue and lock the quote at the expensive one. For example, say Exchange 1 is at NBBO alone with a $10.10 bid and $10.11 offer 100x100. Exchange 1 charges 30 mills per share to remove liquidity. So the price to buy at $10.11 is really equivalent to buying at $10.1130. An algorithm willing to buy 200 shares at $10.11 may choose to post 200 shares at Exchange 2 which offers a fee-fee model at 1 mill per share and thus trying to buy at $10.1101 adjusted for the fee. From a NBBO perspective, it will appear as a locked market (NBB of $10.11, NBO of $10.11 and midpoint of $10.11), on a cost plus basis it is a $10.1101 - $10.1130 market.
Locked markets will thus reduce the spreads adjusted for the access fee and rebate (or make it zero unadjusted for spread) and will increasingly be a norm as opposed to a rule violation. With the liquidity takers having the ability to choose when not to take liquidity rather posting the order at different exchanges at the same price, the power of making choice will shift from liquidity suppliers to liquidity takers. And with the choices shifting so will the pricing of exchanges.
The rebate-funded exchange loses its business model. A venue whose only asset was an unavoidable protected quote now has to answer the question it was never required to answer: why should flow come here? Some will find an answer. Some were never designed to have one.
Move to a fee-fee model from a maker-taker model. Absent the OPR, it is likely that exchanges move to a fee-fee model which will allow for a much lower fee. Note that the SEC has already approved access fee reduction which is slated to go live in November 2027. It is unclear whether that is needed absent OPR. Europe is the closest thing we have to evidence. European markets have no OPR but do have fragmented markets and for most major European exchanges maker-taker is not the prevalent model.
HFT Market Making revision. For the largest HFT firms, who invested heavily in technology and rely on rebate mechanisms for providing liquidity, the dynamics are very likely to change. The rebate was compensation for adverse selection borne on venues sitting last in the routing table. Only the fastest HFT firms benefit from it, those who have built an edge minimizing adverse selection with ultra low latency infrastructure and thus have the ability to retain the rebates. With no obligation to trade at their quoted prices the order flow will likely change. And with fee changes on exchanges, that will change further. How many of those strategies survive and how many get adjusted remains a question. Perhaps it will mean that we see less liquidity on exchanges and thus wider spreads. Or perhaps with reductions in fees, the order flow on exchanges becomes less toxic, leading to more competition between HFTs and thus narrower spreads. Only time will tell and it will likely happen in a sequence of events as opposed to a single change.
The routing hierarchy re-sorts, and ATSs have to re-justify themselves. If taking on an exchange gets meaningfully cheaper, the fee-avoidance ranking that put maker-taker venues last stops being the obvious choice. While that may have been the primary reason for a number of ATSs to originate, now that they are there, it is unclear that repeal does anything to that. Broker dealers will still be incentivized to target their own ATSs for taking liquidity but perhaps there will be less pinging of other ATSs (on the far side) before they route to exchanges if the access fee is reduced on exchanges.
Peg Mid, Peg Near, Peg Far in locked markets. The NBBO is load-bearing far beyond Rule 611. Most of the ATS flow relies on NBBO pegging—whether pegging to the near, far or midpoint price. The repeal of OPR as proposed will lead to extended periods of locked markets with orders pegging near, far and midpoint all pointed to the same price. If we learn from European markets, some ATSs may start referring to the primary market’s BBO and ECNs/exchanges may take a composite price based on the primary market and their own order book—because referring to one market will not lead to locked quotes. Or perhaps some ATSs will use NBBO peg when quotes are locked and the primary market BBO when quotes are not locked. Will this lead to gaming of midpoint orders when the true midpoint is different from the primary midpoint? Could that mean primary markets will increase their market share (as is the case in European markets)? From our experience building and operating algorithms for EMEA and Canadian markets, this is an area that requires a lot of attention from an algo and SOR design and TCA perspective for proper error handling and execution.
We do experience locked markets in US equity markets today but they are largely due to race conditions and disappear quickly. Most ATSs either allow opting out of trading when markets are locked. We opt out, and we pause trading in our algorithms when markets are locked. For other markets (e.g., Canada, EMEA) where occurrences of locked markets and crossed markets are frequent and, more importantly, expected, we use primary market quotes for the period that the quote remains locked or crossed.
As algorithm behavior adjusts, TCA will need to adjust as well. In the presence of locked and crossed markets, what is price improvement? While calculating arrival price do you take the quote prior to a locked or crossed market midpoint? Or simply take primary market midpoint? There will be many questions to answer in order to define a new standard of measurement.
Venues do not need OPR to keep multiplying
While OPR may have directly or indirectly been the reason for the fragmentation in the US equity market, the question is whether its repeal will reduce this fragmentation. Broker-dealers, exchange operators, and market makers have spent years and hundreds of millions of dollars building network effects. They will work hard to maintain their edge and to adjust in light of new regulation and of course they will continue to lobby to shift regulation in their favor. But the more direct evidence that repeal will not consolidate the venue landscape is that the most recent wave of venue creation has had nothing to do with OPR in the first place.
Consider primary listings to start. The Texas Stock Exchange has made its primary differentiation not about trading at all, but about using Texas's corporate law. Texas’s corporate law is favorable to the issuer and “pro business” even if a company is dual listed. NYSE and Nasdaq have responded by rebranding their inverted exchanges, which had low volume to begin with, and reincorporating them in Texas as NYSE Texas and Nasdaq Texas. For another example, the exchange 24X launched with its differentiation defined by trading 24 hours a day, and Nasdaq followed by acquiring Level ATS to help it expand into 24-hour trading. And finally, there is also considerable buzz about security tokenization, rumored to be a motivation behind OPR’s repeal rather than fragmentation (since repeal of OPR is a necessary condition for operating a tokenized exchange). And this could be where the next wave of new venues actually comes from.
So, what about the genie?
The impact of regulation as important as Reg NMS is never immediate. It takes its own course, through a chain reaction of adjustments, and the second-order effects are usually larger than the first-order ones the rule was written to produce. Reg NMS with OPR has arguably been the most influential regulation in the largest securities market in the world.
Its repeal will perhaps be even more consequential because the last two decades of infrastructure have been built around it. There will be some good and intended effects—exchanges will be forced to answer what value they bring to the table beyond simply offering a protected quote from a fleeting market maker. But there are likely to be unintended effects unfolding as well.
Take European regulation MiFID as as an example. The European Union’s MiFID II framework intended to curb dark trading and force transactions onto public, transparent exchanges, but it unintendedly triggered a fragmentation of liquidity as traders shifted from traditional dark pools into alternative, opaque venues like dealer-owned systematic internalizers. Both anticipated and unanticipated effects have emerged.
While the bar to start a new exchange will surely be higher, no retail broker or institutional broker trades only on exchanges. The fragmentation outside the exchanges is at least an order of magnitude higher than on exchanges. The venues that already exist will continue to defend the network effects they have already built and perhaps increase their territories into other areas such as 24-hour trading, tokenization, and other business models we can only imagine after the immediate reaction from the regulation takes hold. For this reason, I think fragmentation is more likely to change its form than to go away—the genie finding a new home outside the bottle, if you will—though cited as the primary objective for repeal.
Twenty years ago the SEC took a bold step to tie the markets together with Reg NMS and has spent the last twenty years fixing the imperfect aspects of the market one by one, never sacrificing the sanctity of the NBBO. Changes have included adding Self Help rules and testing procedures to ensure that one bad exchange cannot bring down the ecosystem, improving the quality of the NBBO by reducing odd lots for highly priced securities, improving the latency of SIP, reducing the access fee and thus lowering the cost-adjusted spread, reducing the tick size for liquid stocks. All of these changes were measured and deliberate. Repeal of the OPR will be far reaching, perhaps in both intended and unintended ways.
At BestEx Research, our answer to fragmentation has never depended on the protected quote, and it will not depend on its absence. We will continue to observe and act as market structure evolves, adjusting our strategies and keeping you abreast of our findings.
------------------------
Footnotes:
1SEC has proposed rescinding Rule 611 (the Order Protection Rule or trade-through rule) and Rule 610(e) (the prohibition on locked and crossed markets) under Regulation NMS. They are related—Rule 611 prohibits exchanges from trading at a worse price than another exchange displaying a better price, and locked and crossed markets prevent them from posting a price if liquidity at the same price or better is available at another exchange. I refer to these rules together for the rest of the paper.
2Rule 612, as part of Reg NMS, banned quoting in increments of less than $.01 for stocks with price >$1.
330 mills of access fee in a roundtrip is equivalent to 60 mills, which on a notional price of $5 is equivalent to .12% or 12 basis points.
