You have been granted access to this page through First Click Free. Subsequent use of TabbFORUM will require logging in. If you don't have an account, registration is free.

Videos

  • Rail_thumb_chicago

    The HFT Arms Race

    Eric Budish, Associate Professor of Economics at the University of Chicago Booth School of Business, discusses the arms race between high-frequency traders and whether the ceaseless drive for faster connection ...
     
  • Rail_thumb_o_brien

    Bats' O'Brien on Market Role of HFT, Dark Pools

    Bill O'Brien, president of Bats Global Markets, talks about the role of high-frequency trading and dark pools in financial markets and the history of Reg NMs. He speaks with Erik Schatzker on Bloomberg ...
     
  • Rail_thumb_brad

    IEX’s Katsuyama: What I meant by 'rigged'

    Does IEX president & CEO Brad Katsuyama regret saying the markets are "rigged"? Katsuyama explains what he really meant on CNBC’s Fast Money.   Related video: "Michael Lewis ...
     
 

More Video | Podcasts

Advertisement
Spotlight-black
The Great HFT Debate

10 March 2014

The Order Type Controversy, Part I: Price to Comply

Since the order type controversy first gained attention two years ago, regulatory pressure has brought change to exchange order matching practices, while the HFT firms that once dominated the marketplace have waned. But we have not yet seen the adoption of advanced order types by a ‘smarter’ buy side. To better understand and identify appropriate usage of complex order types, it helps to understand how they evolved since the introduction of Regulation NMS in 2007.

Two years have passed since what is now known as the “order type controversy” first hit the press. Over this period we have seen regulatory scrutiny directly impact exchange order matching practices, resulting in improved disclosure and transparency at major exchanges, and the waning of HFT firms that once dominated the marketplace. Meanwhile, leading firms such as NYSE Euronext, as well as new entrants such as IEX, have made visible steps to level the playing field among participants. While these developments are generally positive, however, we have not yet seen the wholesale adoption of advanced order types by a “smarter” buy side that can confidently navigate the new electronic marketplace.

To better understand and identify appropriate usage of complex order types, it helps to understand the history and context of how order types evolved since the implementation of Regulation NMS in 2007, while also assessing the different implementations across exchanges. In this series, I focus primarily on orders that “hide and light” and their corresponding price-sliding practices.

[View more of Haim Bodek's TabbFORUM commentaries on order types, high-frequency trading and market structure.]

For better or worse, the credit for inventing the first “hide and light” order type should be given to NASDAQ for its Price to Comply order type. Originally conceived in mid-2006, the Price to Comply order type was intended to assist NASDAQ in integrating its three platforms into a single order book in a manner that respected the recently adopted Regulation NMS framework.

The influence of the Price to Comply order would find its way into every exchange in the US equities marketplace. It was an innovation that would permanently conjoin the high-frequency trading tradition with the nuances of Regulation NMS, and in particular with Rule 610(d), which banned locked markets.

As I have described in previous articles, Regulation NMS was, among other things, a regulatory framework that was intended to bind the highly fragmented US equities marketplace into a coherent national best bid and offer (NBBO). One of the problems it sought to address was the problem of locked and crossed markets, where the best price on one securities exchange was marketable against the best price on a competing venue. The concern was that a locked market was perceived as a manifestation of inefficiency, an aberration of a national market system that was intended to match parties willing to trade at the same price when their offsetting orders were represented at competing market venues.

However, in part the concern was also aimed at intentional distortions of the maker-taker model that made it economically sensible for parties to lock markets to attempt to execute for rebate, when such parties would otherwise incur a taker fee if they were routed to the venue displaying the best price. In the words of regulators: “Often, the locking market participant is not truly willing to trade at the displayed locking price, but instead chooses to lock rather than execute against the already-displayed quotation to receive a liquidity rebate.” Thus Rule 610(d) of Regulation NMS and the ban of locked markets were also directed at eliminating one of the primary activities of high-frequency traders engaged in rebate-subsidized scalping strategies. Clearly, Rule 610(d) was unsuccessful within this regard.

To grasp the significance of the Price to Comply order type, we first need to focus on its stated purpose: to re-price an order to comply with Reg NMS Rule 610(d). By this, the rule stipulates that NASDAQ was not to display a quote that was the best bid if that price was equal or greater to the best offer on an away exchange – say, NYSE ARCA (and comparable logic applies for the case of a displayed offer). The rule also emphasized the obligation of NASDAQ to proactively police its members and to “prohibit its members from engaging in a pattern or practice of displaying quotations that lock or cross any protected quotation in an NMS stock.”

Now even before we review the details of how Price to Comply in fact disallows the locked market condition above, it is sensible to question why a complex order type is required at all to comply with Regulation NMS. And the short answer is that Price to Comply is thoroughly unnecessary for complying with Rule 610(d) and the ban on locked markets. In actuality, Price to Comply is an optional tool that provides sophisticated market participants with a way to address the inconvenience of complying with Rule 610(d) in a variety of scenarios where such participants are incentivized to lock markets.

[Related "Take the Time to Understand the Complexities of the Markets"]

Now clearly, the simplest way to address a locked market would be to reroute the order to the away exchange for execution. However, the problem with this simple solution is that it interferes with the primary interests of the exchange to encourage aggressive posting that results in captive volume on the venue. Furthermore, the solution of re-routing an order to an away exchange also interferes with the primary functions of smart-order-routers and high-frequency traders, both of which have no incentive to use such routable orders.

Consider an executing broker that has invested in smart order routing technology. The idea of paying an execution fee for using NASDAQ’ssmart order router to access an away exchange is unacceptable. The incentive for such a broker is to cancel and reroute the order through its own pipes, thus avoiding the redundant routing fee while fulfilling its best execution duty.

Likewise, consider a high-frequency trader that is posting aggressively into a locked market, aiming to collect a rebate in a liquidity-providing function (and let us forget for the moment that such activity is in fact prohibited by rule 610(d) as a pattern or practice of locking away quotations). For the HFT, the idea of having his aggressive orders re-routed unexpectedly, thus incurring both a routing fee and a taker fee when executed at the away market, is completely unacceptable. And this discussion doesn’t even address the opportunity cost the HFT bears when his routed order loses the opportunity to gain a favorable top-of-queue position and collect a rebate.

So it is not only in the interest of the exchange to retain volume by providing an alternative to routable orders, it also is in the interest of smart order routers and high-frequency traders. Both types of sophisticated market participants require NASDAQ to provide non-routable orders compliant with rule 610(d) to achieve sensible economics.

The simplest way to implement a non-routable order that complies with the ban on locked markets is to kick back the order to the customer as impermissible when an away market is locked. Another solution is to adjust a non-routable order that locks a market by ticking it a price back (to the back of the queue), which NASDAQ in fact supported with its Price to Display order. However, clearly both options are unattractive for an exchange incentivized to promote volume by retaining aggressive orders that are likely to result in executions. NASDAQ’s solution was Price to Comply, the innovative order type that achieved “compliance” with Rule 610(d) by cleverly exploiting a legal exception implicit in the Regulation NMS framework itself to allow orders to be accepted at the locking price.

Price to Comply leveraged an artifact of the Regulation NMS framework that distinguished between protected quotations and hidden orders in order to “comply” with Rule 610(d). Under this rule, one wasn’t allowed to lock a protected quotation by displaying a price that locked the market, but hidden orders were exempt from the ban. Under the regulation, hidden orders were not considered protected quotations and thus could freely be used to lock markets. In other words, Price to Comply offered the user the capability of locking a market with an exempt hidden order in the cases where his displayed order would violate the ban on locked markets. 

When NASDAQ received a Price to Comply order, it would check if the order price locked an away market. If the order was in fact impermissible according to Rule 610(d) and the ban on locked markets, NASDAQ would book the Price to Comply order as a hidden order at the locking price. Price to Comply was the “have your cake and eat it too” order, conveniently booked as displayed or hidden – whichever was more preferable when considered in the context of a permissible display price according to Rule 610(d).

Price to Comply had another interesting property: It would be canceled back to the user if the away market at any point was displayed at a price that was through the hidden price, which would otherwise result in a trade-through violation prohibited by Rule 611 of Regulation NMS. As we shall see later in the series, such secondary features can have significant impact on execution quality, especially when order type interaction is taken into account.

Lastly, “book once” and re-bookable versions of the order type were available depending on order interface, a topic that will be examined in more detail in the next article.

Overall, Price to Comply appears at first glance to be a relatively unsophisticated order that is no different from a traditional limit order – except in the event where it locks an away market upon order receipt in the matching engine. In that critical scenario, the “book once” version of Price to Comply is booked as a hidden order, remaining in that state until it is either executed at the hidden price or canceled back to the user in the specific case where an away market crosses through its price.

However, inherent in the formulation of the Price to Comply order type is a deep resistance to the inconvenience of Rule 610(d). The ban on locked markets simply interfered with the economic incentives of sophisticated sell-side brokers, high-frequency traders, and NASDAQ itself to permit the booking of aggressive orders that locked markets. Thus Price to Comply – the first “hide and light” order – was introduced to employ the clever regulatory workaround of a hidden state to comply with Rule 610(d) and the ban on locked markets.

In fact, when NASDAQ announced in April 2007 that it would expand the usage of the Price to Comply feature set by mandating that all variants of non-routable order types not marked InterMarket Sweep Order (ISO) employ the hidden state when joining a locked market, NASDAQ further confirmed its intention to center its order matching engine on practices that were designed primarily to assist in effectively locking markets for all practical purposes.

Price to Comply set a critical precedent for the industry as a whole where the use of convoluted order handling logic, regulatory workarounds, and opaque feature sets provided a convenient mechanism to address the inconvenience of Regulation NMS in a “compliant” manner. Indeed, with hindsight as our guide, Price to Comply might more aptly have been named “Price to Circumvent.”

In my next article, I journey down the market structure rabbit hole to examine features NASDAQ embedded in the Price to Comply order type to enable “lighting” and how such features differ between FIX and OUCH protocols.

Spotlight-white-trans For more stories in the The Great HFT Debate Spotlight Series click here.

Comments | Post a Comment

8 Comments to "The Order Type Controversy, Part I: Price to Comply":
  • Missing
    rbattali

    10 March 2014

    Nice presentation - will use it in my trading and markets class. 

    The increased alpha generated by executing at more favorable prices more than compensates investors for the lower fill rates and the lower maker rebates associated with the use of this order type.

  • Missing
    JCD

    11 March 2014

    Well written piece describing the origins of the dysfunctions that Reg NMS foisted on the equity markets. Looking forward to the next installment.

  • Comment_dave_cummings
    Dave Cummings

    13 March 2014

    "Price to Comply" orders are designed specifically to comply exactly with Reg NMS.  They are the most aggressive price allowed without removing liquidity, which as Haim points out, costs additional fees.  

    If you do not allow this order type, it results in needless 'machine-gunning' orders which get rejected until the exchange gets the data the the away market has lifted.  

    When you look at it closely, this order type makes sense.  It is available to all participants.  It is well documented.  There is nothing unfair about it as some people have claimed.

  • Comment_haim_linkedin
    hbodek

    13 March 2014

    It is hard to rationalize Dave's comments with Price to Comply in particular which in fact encouraged HFT machine gunning strategies on NASDAQ from 2007-2012. I will address this oddity in the next article which details the differences between OUCH and FIX versions of Price to Comply.

    Dave's comments, while appreciated, are more representative of what "hide and light" orders could have been had such orders focused on the needs of buy-side traders and not HFTs. Later in the series I will review changes NASDAQ made to such order types in 2011-2012 which in my opinion set the standard for making investor-friendly versions of these order types that do not accommodate HFT strategies.

  • Comment_tdxme
    czarcu

    18 March 2014

    Haim -- the existence of these specialized order types created to "comply" with Reg NMS, would suggest that the regulators were successful in their implementation of Reg 610(d), and have indeed eliminated Locked and Crossed markets.  Yet both are common occurrences in today's market place. Furthermore, careful study of Locked markets reveals very well organized patterns of behavior in many cases, seemingly designed not only to lock the market, but also to push the NBBO in a desired direction.  So I would argue that the existence of advanced order types turbocharges the previous "rebate-subsidized scalping strategies", making them even more profitable.  It seems to me that exchanges need to take a closer look at how they handle orders that do indeed lock markets, as they are the first defense against that...  

  • Comment_haim_linkedin
    hbodek

    19 March 2014

    Cristian -- I appreciate your comments. You are right in recognizing that "lighting" events can have significant market impact. In the process of constraining permissible displayed prices, batches of "hide and light" orders that are displayed in a single "lighting" event can create a self-fulfilling prophecy in short-term market-microstructure momentum. This is especially the case with advanced versions of "hide and light" orders that are designed to light up multiple times at increasingly aggressive prices. 

    As you indicate, I do indeed claim that such order types supercharge HFT scalping strategies and we are dependent on exchanges to fulfill their SRO duties to prevent discriminatory practices from unfairly disadvantaging the public customer. I would like to add that "hide and light" order types become particularly compelling when you combine them with post-only features that assist in capturing rebate from less sophisticated market participants. I will of course cover this topic as the series progresses. 

    Haim

  • Comment_tdxme
    czarcu

    20 March 2014

    Thanks Haim...I'm not convinced that exchanges have the capability to recognize that scalping behavior even occurs, let alone do something about it.   They can't even enforce their own rules -- they all claim to be able to prevent markets from locking, yet evidence suggests the exact opposite!  I see markets locked by orders originated by all the major exchanges, no exceptions...pretty sad state of affairs to be exact! 

  • Anon_avatar
    Anonymous

    15 April 2014

    Haim: Thanks. Wonderful, incisive review. The SEC and the SROs have been a big part of the problem here. One wonders what awaits the latter as the SEC, for whatever reason, finally stirs from its regulatory paralysis. Whether or not it confesses error for having allowed all this to occur in the first place, the need for reformation of its market structure rules seems clear and compelling. The agency needs to stop hiding behind the skirts of a purported need for more and better economic analysis before doing anything and using that as a continuimg excuse for inaction. Can class action litigators help all this along? Maybe, maybe not. After all, the SEC was actively engaged in overseeing all this as it happened (talk about eyes wide shut!) -- and that's probably a stopper as far as the exchanges are concerned.  Remember what happened to those who years ago thought payment for order flow was unlawful under state bribery statutes? That didn't work out too well, did it? In any event, I await your continuation in Part 2 with interest.

You must log in to comment.