When was the last “Air Raid”? (Or: Why wait to eliminate the Order Protection Rule?)

It has often been said that one of the greatest impediments to innovation and positive change is the expression “It has always been done this way.”    Such thinking is pernicious, as people assume that “this way” must be correct and that change must imply risk, particularly when the people involved either forget or ignore WHY things are done that way…

To illustrate, consider the following story told to creativity expert Michael Michalko:

A quality management consultant was hired by a small English manufacturing company to advise them on improving general operating efficiency. The company produced a report which dealt with various aspects of productivity. At the top-right corner of one form, there was a small box. The consultant noted that the figure ‘0’ had been written in every such report for the past year. On questioning the members of the staff who completed the report, they told him that they always put a zero in that box, and when he asked them why they told him they were told do so by their supervisor. The supervisor told him he guessed it had to do with accidents but wasn’t sure. It had always been “0” for the twenty-five years he had been there, so he continued the practice. It, too, was something he was told to do by his former supervisor. 

The consultant could find no one in the company who could tell him what the box represented. Intrigued, he went to the warehouse where the company kept its archives to see what he could discover about the form. The company was founded in 1937 and the records were preserved all the way back to 1940. He found the old reports, he saw that the zero return had continued uninterrupted for as far back as the records extended. Eventually, he found the box that catalogued all the originals of the forms the company had used during its history dating back to 1940. In it, he found the original report which was created in 1941, in pristine condition. In the top right corner was the mysterious box, with the heading clearly shown …… ‘Number of Air Raids Today.’  Over time, the heading disappeared but the box remained.[1]


This tendency has impacted the current debate over Regulation NMS.  While Reg NMS did usher in a wave of technological innovation, it is important to review the historical context, which clearly shows that, at a minimum the Order Protection Rule (OPR) and ban on locked markets are no longer needed.

The market structure before Reg NMS was bifurcated.   Nasdaq stocks traded electronically and stocks listed on the New York Stock Exchange (NYSE) traded predominantly on their trading floor.   The NYSE maintained over 80% market share and efforts to compete by electronic markets and regional stock exchanges were repeatedly frustrated.   This was not accidental, there were artificial barriers created by the NYSE’s manual trading model and rules.

Nasdaq, whose system had become fully electronic at the end of the 90s, faced intense competition from alternative trading systems (ATSs).   The growth of ATSs accelerated once Nasdaq went fully electronic, fueled, in part, by the development of Smart Order Routers (SORs) that aggregated liquidity for traders across all electronically accessible markets that published quote feeds.  There were over a dozen ATSs formed, and, contrary to today’s landscape where ATSs are all “dark”, in those days, most displayed visible quotes based on their order books.    (ATSs that display quotes are called Electronic Communication Networks or ECNs).   M&A activity in the ECN space was intense, however, with Nasdaq buying two of its main competitors (Instinet, which had earlier bought Island forming the largest ECN, and Brut, which had earlier merged with Strike, and was purchased mainly for its routing broker).  Nasdaq’s purchasing spree enhanced fears of a “duopoly” consisting of Nasdaq and NYSE, which, brokers feared, would lead to rising costs.   That fear, in turn, led to the formation of both BATS by a consortium led by Tradebot’s CEO David Cummings and DirectEdge, which was formed when a Knight Securities led consortium bought and re-platformed the Attain ECN.    Around the same time, Archipelago bought the Pacific Stock Exchange after having earlier merged with Redibook to form another large ECN.  Archipelago was then purchased by the NYSE in a reverse takeover to form NYSE Arca in early 2006, inspired, at least in part, by the passage of Reg NMS.

The NYSE, before the implementation of Reg NMS, operated a floor based system which required specialist intervention to execute all trades.  It did support some automation, such as handheld devices on the floor and the Designated Order Turnaround (DOT) system, but those systems still needed specialist approval to execute orders, even when matched.  This made it difficult for SORs to interact with the NYSE and ECNs at the same time, since the NYSE could take a long time to respond to orders.    In such an environment, it was hard to argue that institutions were better served by trading away from the NYSE, as most of the trading took place there.  The only official system connecting the floor and competing (regional) exchanges was called ITS (Intermarket Trading System), which was slow and not suited for modern SORs.  Regional exchanges complained bitterly that the specialists in NY ignored their quotes, despite ITS, making it difficult for them to satisfy their clients.   Meanwhile, ECN attempts to compete with the NYSE were frustrated by rules such as NYSE rule 390, which prohibited member firms from most off exchange trading and by the fact that there were no fully executable quotes from the NYSE to compare ECN quotes to.

This environment let to the creation of an unlikely coalition of regional exchanges, ECN owners, electronic trading firms and large broker dealers who asked the SEC to promote automation and break the NYSE’s virtual monopoly.  The resulting compromise was to create the Order Protection Rule (OPR) for top-of-book quotes and the prohibition on locked markets within Regulation NMS.  (There was also a new formula for apportioning market data fees, a ban on sub-penny quotes, and a cap on access fees, but those are subjects for different notes.)  The OPR was designed to force the NYSE to make their quotes fully and electronically executable to qualify for order protection.  It also forced routing firms to access quotes from away markets before trading against their own book at inferior prices.  The prohibition against locked markets, meanwhile, was promulgated because of the “duopoly” fear.  Regional exchanges, ECNs, and their users, were concerned that the NYSE specialists or large Nasdaq market makers would ignore their quotes and choose to lock them instead of routing to their systems.  Since most of the markets involved were “maker/taker”, this behavior would be costly to these markets since the trader who posted was paid a rebate, while routing to access liquidity incurred an access fee.    This difference averaged over a half cent per share, which was (is) quite meaningful.

Today, however, the market has changed and the valid concerns of last decade are mostly irrelevant.  First, no exchange has monopoly power anymore, and it is extremely unlikely that non-automated exchange systems would attract enough order flow to pose a problem.  Second, rules on best execution require brokers to access all displayed liquidity unless the brokers can document reasons why doing so would harm their client.  FINRA has explicitly stated that failing to execute a client order because a broker wanted to control their costs would be judged a violation of best execution[2].  Most participants believe that these rules also apply to exchange routing, although I will admit that codifying that would be beneficial.

Therefore, there is no reason for market participants to have the same fears it did before Reg NMS was implemented.  Not only is that fear misplaced, but keeping the OPR and locked market prohibition directly harms investors by increased complexity, wider spreads, and less innovation.  The increased complexity caused by these rules has manifested in many new order types, price sliding mechanisms, and message traffic.   New order types included many variations of Intermarket Sweep Orders (ISOs), which were explicitly created by the OPR, while price sliding was created by exchanges to avoid having to cancel orders that violate the rules.  While there are many variations, price sliding is an exchange service that re-prices a limit order to avoid locking another market, and then, in certain circumstances, re-prices them again if the other market moves.  The result of these orders and services, along with competition to predict when an exchanges price is going to catch up with other exchanges, was that message traffic increased substantially.   It is, therefore, reasonable, to expect a decrease in message traffic if the OPR and locked market ban were repealed.

Eliminating locked markets directly increases the average spread costs paid by investors. For retail investor, locked markets are a “choice” price as they get to choose whether to buy or sell at the same price, and do not pay access fees to exchanges[3].  While locked markets are not truly a “choice” price for institutional investors, due to fees, the net spread is much smaller than the minimum one penny increment.

Lastly, the OPR and locked market prohibition contribute to suppressing innovation.  Block trading venues, auction markets, and various other market models are more difficult to implement due to these rules.   While best execution rules would still be operative, meaning that venues that ignore significant liquidity will likely attract scrutiny, eliminating the OPR would facilitate market models that match buyers and sellers within a single system.   Under the OPR, such systems must either send orders to all displayed quotes and wait for responses, before printing trades outside of the NBBO, or ignore situations where buyers and sellers don’t match up at that price.   In addition, everyone’s “favorite” market structure bogeyman, the speed-bump, would be less contentious if participants were not required to route to venues that have them.[4]

Unfortunately, even the critics of Regulation NMS feel the need to preserve these rules, while a “holistic review” is undertaken.   In my opinion, that is wrong.  While a holistic review of market structure is warranted, de-regulation by eliminating both the OPR and the ban on locked markets makes obvious sense.  The markets, and the regulatory environment have changed sufficiently to warrant immediate action.  A holistic review, if history is a guide, will take many YEARS to complete, and therefore, should not be used as a delaying tactic to maintain the status quo.   In short, the last air raid is now a distant memory, and there is no need to protect exchange quotes due to the fear of dropping bombs from airplanes that no longer exist…


[1] http://www.creativitypost.com/business/its_always_done_this_way,  Michael Michalko, Oct 1, 2012

[2] FINRA Regulatory Notice 15-46 https://www.finra.org/sites/default/files/notice_doc_file_ref/Notice_Regulatory_15-46.pdf

[3] Marketable retail investor orders, according to Rule 606 data are mostly sent to wholesale market makers for no cost to the retail broker.  According to 605 statistics, such market makers, on average execute those orders inside of the bid offer spread.

[4] The other issue with speedbumps is that venues who slow quotes down would still be included in the SIP NBBO used for best execution reporting.  Without quote protection, however, there would likely be less resistance to limiting NBBO inclusion to quotes that have no delays, as exclusion would have less of an economic consequence.

One thought on “When was the last “Air Raid”? (Or: Why wait to eliminate the Order Protection Rule?)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s