Subject: File No. SR-NYSE-2004-05
From: George Rutherfurd, Consultant
March 10, 2005
(The author has been a consultant to several non-U.S. securities markets in their transition from floor-based to screen-based trading. The author is writing on behalfof two institutional investing organisations that wish to remain anonymous at this time. The author wishes to acknowledge the assistance given to him over the years by the SEC staff in understanding both U.S. federal and SRO regulations, as well as assistance over the years from several NYSE specialists and floor brokers)
March 10, 2005
As discussed in greater detail in the body of this letter, the NYSE's proposal emphasises several of the worst aspects of its current floor market in its lack of transparency, and in its conferring advantages to floor brokers and specialists which are denied to the general investing public. Far from protecting the public limit order book, the NYSE has proposed to give public orders entering its market executions which, in many instances, would be inferior to what those orders would receive in the current floor auction, and is proposing to permit specialists to represent a clearly illegal order type. The NYSE proposal is suffused with ludicrous jargon that obscures (perhaps intentionally) rather than clarifies how the proposal would actually operate.
Amendment No. 2 to SR-NYSE-2004-05 repaired some, but by no means all, of the damage created by Amendment No. 1, easily the most confusing U.S. SRO rule submission I've looked at in the past 20 years. The entire proposal reads as though it were simply thrown together by floor traders and then waved out of the building by the NYSE's legal staff, which doubtless did not really understand it.
THE NYSE MUST RE-CAST ITS PROPOSAL IN PLAIN ENGLISH
The NYSE's proposal is riddled with self-invented jargon that does not fairly and accurately describe exactly what the NYSE intends. This is not simply an issue of editorial comment or drafting preference. I am referring to the fundamental meaning of words here, and words, and word choices, matter greatly if a proposal is to be easily comprehended by the NYSE's user community.
The SEC itself is subject to so-called "plain English" requirements. It is simply a matter of basic public interest for the SEC to hold the NYSE to "plain English" standards as well.
The following terms in particular are ridiculously obfuscatory and do not express at all what the NYSE really means.
- "Auction limit order." If there were such a thing as a counter-intuitive hall of fame, this term would be a shoo-in for inclusion. Limit orders on the NYSE are fully subject to execution, often over time, in the NYSE's auction. The so-called "auction limit order" (a marketable limit order that will be briefly displayed for price improvement) has only a fleeting, tangential connection to the auction, and most of the time would be executed automatically, and not even in the floor auction. Yet this order is to be denominated "auction", and not the limit orders that are actually auction-based. What the NYSE really means here is "improved price limit order" and the SEC should insist that this plain English term (or a similar, simple descriptive term) be used.
- "Liquidity replenishment point." This comically bizarre formulation gives the reader no clue as to what the NYSE really means. The NYSE is merely talking about a kind of volatility surge protector, which shuts off automatic executions and autoquoting for 5 to 10 seconds. The basic concept is not unsound, but it is simply an arbitrary time-out, based on pre-determined parameters that have nothing to do with the actual liquidity situtation at the time the surge protector kicks in. The 5 to 10 second time periods are so brief that obviously there is not going to be a whole lot of carefully considered "replenishment" going on (nor, since the parameters are pre-set in any case and not based on actual market conditions, would there necessarily be any need for "replenishment.") What the NYSE really means here is simply "auto ex shut-off point", and the SEC should insist that this plain English term (or a similar, simple descriptive term) be used.
- "Broker agency interest file." This seemingly benign gibberish inartfully disguises the self-serving heart and soul of the purported "hybrid market", namely the creation of what is, in effect, a hidden, second, automated limit order book (to which floor brokers alone may input orders), which will compete directly with, and siphon executions away from, the conventional public limit order book. The Securities Exchange Act, as well as SEC and NYSE rules, use the term "order;" the term "agency interest" is of no legal cognisance or consequence, and should not be used, particularly in rule text. The NYSE has attempted to mask what is really happening by talking about "agency interest" that "arrives" into a "file." But what is actually happening is that brokers' orders (not "agency interest") are being entered (orders do not "arrive") onto a limit order book (not a "file"). No matter how many separate brokers' files there may be, there is only one overall cyberspace re pository for them, and this has to be called a limit order book, the term understood in the securities industry for a repository of orders awaiting execution. This second, hidden limit order book is proposed to trade on "parity" with the conventional public limit order book. "Parity" is the NYSE's jargon for its unique form of side-by-side trading whereby floor brokers with "go along" orders split executions with each other, and with the public limit order book, effectively eroding the time of order entry priority of orders on the public limit order book. What the NYSE really means here is "go along limit order book", and the SEC should insist that this plain English term (or a similar, simple descriptive term) be used.
- "Specialist interest file." This gibberish term disguises the proposed creation of a third, hidden automated limit order book, which would contain the specialist's dealer orders. This book could not compete at the same price with the public limit order book (the specialist is the agent for those orders) but can split executions with the go along limit order book if the specialist is liquidating or decreasing a position. What the NYSE really means here is "specialist dealer order book', and the SEC should insist that this plain English term (or a similar, simple descriptive term) be used.
- "Residual." This is a confusing term that has far from universal currency with investors. The NYSE is merely referring to the "remaining balance" of an auto ex order after it has been executed against the displayed bid or offer, or after it has been "cleaned up" to its limit price (or one of the other technical exceptions applies). As the NYSE is simply referring to the "remaining balance" of an order, the SEC should insist that this plain English term (or a similar, simple descriptive term) be used.
OVERVIEW OF THE PROPOSED "HYBRID MARKET" (USING THE PLAIN ENGLISH TERMS NOTED ABOVE)
Notwithstanding all its marketing hyperbole, the NYSE's "hybrid market" proposal is actually a rather modest affair. The NYSE is proposing to expand its Direct (automatic execution) system by eliminating order entry size limitations and by providing that "auto ex" orders (non-auction orders to be executed by computer), after they have exhausted the NYSE's published bid or offer, will be "swept" to a single "clean- up" price, which is the lowest price (in the case of a sell order) or the highest price (in the case of a buy order) at which contra side liquidity in the floor market ( the three different limit order books noted above) can fill the entire remaining balance of the order within the limit price of the order. (There are several other technical exceptions here).
The NYSE is proposing that improved price limit orders and non-auto ex market orders be displayed for as long as 15 seconds (to obtain possible price improvement) before being executed against contra side interest, which could be at a worse price than existed when the orders entered the market.
The NYSE is proposng to adopt two new, non-transparent, automated limit order books, which will operate in parallel with the traditional public limit order book. The first new limit order book would contain go along (parity) orders that only a floor broker would be permitted to enter, provided that the subject security is traded at one of five contiguous "panels" where the broker is physically situated. The second new limit order book would contain away from the market limit orders entered by the specialist on behalf of the specialist's dealer account.
The NYSE is also proposing to permit specialists (but no one else) to engage in pre-programmed "algorithmic" trading against incoming order flow (ostensibly to provide price improvement), as well as to supplement existing liquidity and re-quote the market following a trade.
"GO ALONG" TRADING IS PREMISED ON THE ANTIQUATED FICTION THAT "A TRADE CLEARS THE FLOOR"
Under NYSE Rule 72, the first bid or offer made at a particular price has priority for execution in the next trade at that price. Once a trade takes place at that price, all non-priority bids and offers at that price are deemed to be re-entered on "parity", meaning that they split executions among those bidding and offering. The NYSE refers to this process of eliminating the time priority based on time of actual order entry as "a trade clears the floor" of such time priority. "Parity" is based on bids and offers, not orders. The public limit order book may contain many limit orders at a particular price, but the specialist makes only one bid on behalf of the book. Each individual floor broker, who may have only one order, is also permitted to make one bid. In active stocks, a trade for as little as 100 shares erodes the time priority of limit orders on the public limit order book, orders which may have been entered well prior in time to a floor broker's even showing up i n a trading crowd. For all practical purposes, the public limit order book is forced into unequal splits with floor brokers representing go along orders all day long. For example, if the specialist is making one bid on behalf of five orders on the public limit order book, and four late arriving floor brokers are each making one go along bid on behalf of a single order each is representing, the public limit order book will receive only 20 percent of a trade, and each floor broker will receive 20 percent, even though the floor broker orders may have been entered later in time than the orders on the public limit order book. If the market then moves away from the limit prices of the orders on the public limit order book, they will go unfilled (because they have to split executions, rather than be executed in time priority order) even though they were entered prior in time to the floor broker go along orders. It is little wonder that the NYSE is fiercely opposed to the "depth of book " alternative that the SEC is proposing in its national market system release; the NYSE does not even fully protect the transparent, fully disclosed public limit orders on its own public limit order books.
Rule 72 also provides a concept called "precedence", meaning that a bid or offer that can trade with the entire size of an incoming order may supersede the lesser-sized go along orders or the public limit order book (which itself can also claim precedence). In practice, NYSE specialists tell me, precedence is rarely, if ever, invoked, as the "crowd culture" is to split executions equally among all go along traders. (A specialist once told me he would be "killed in the ratings" if he invoked precedence, an apparent reference to the periodic judgments floor brokers make of specialists). This is the real specialist trading scandal on the NYSE, and the NYSE has ignored it, and the SEC has missed it. Specialists routinely fail to exercise their fiduciary duty on behalf of the public limit order book to claim precedence to ensure that the orders the represent as agent are in fact executed. The public harm here is far greater than he relative "pocket change" involved in the current specialist trading scandal.
HOW GO ALONG TRADING SHOULD BE CONDUCTED
There is nothing at all inherently wrong with go along trading; it is a conventional strategy and should be facilitated in the marketplace, but not at the expense of a fully transparent public limit order book. At any given price level, orders on the public limit order book should be executed first, in order of time of entry. Then, the go along orders should be permtted to split executions among themselves. The basic concept is simple and fair: transparent, fully disclosed orders should trade ahead of hidden orders.
There is no real disadvantage here to go along investors, as they can always choose to place an order on the public limit order book. Such investors should not, however, be permitted to circumvent the price/time priority of the public limit order book by entering hidden go along orders.
The NYSE's outmoded concepts of "parity" and "a trade clears the floor" may have made sense a century or so ago when they were adopted (I've heard different time periods from the NYSE staff over the years) and stocks traded largely by appointment. Today, however, "parity" is simply a euphemism for "no price/time priority" in active stocks most of the time, a position unknown in any other major securities market.
The SEC needs to insist upon the price/time priority of fully disclosed orders on the public limit order book viv-a-vis hidden go along orders. Alternatively, the SEC should insist that specialists be required to claim precedence at all times, and that precedence be accorded to the public limit order book in all "sweep" transactions.
THE PROPOSED GO ALONG LIMIT ORDER BOOK GIVES FLOOR BROKERS AN EVEN GREATER ADVANTAGE OVER THE PUBLIC LIMIT ORDER BOOK THAN THEY ENJOY TODAY
The proposed electronic go along limit order book involves a radical redefinition of "parity" and is not at all the functional equivalent of go along trading in the current floor auction. There is no such concept in Rule 72 as away from the market parity of bids and offers. At each price change, new bids and offers are made, with the first bid/offer at that new price level having time priority. Specialists tell me that, typically, they make a new bid or offer on behalf of the public limit order book, which then has priority (at least until the next trade) over floor broker go along orders. This process keeps repeating itself as price levels change, and is fully transparent, at least to those at the point of sale.
What the NYSE is proposing, however, in its hidden, go along limit order book, is that floor brokers may enter firm, away from the market go along orders that are "hard wired" to trade automatically, which is not at all the way floor brokers are required to bid and offer, at each price level as that price level is reached, in the floor auction.
There are at least two significant harms in what the NYSE is proposing:
- The price/time priority of the public limit order book, already severely compromised in the conventional floor auction, is further obliterated in "sweep" transactions, even though the orders on the public limit order book could probably claim time priority in the NYSE's current floor auction clean up process.
- Floor brokers would be given a unique and unfair advantage over the public. They can see the orders on the public limit order book, but investors who have entered, or are thinking of entering, public limit orders cannot see the hidden go along orders, which floor brokers can enter with a kind of "insider" knowledge as to the best prices at which to interact with contra side order flow and supersede (or neuter) the public limit order book's price/time priority. Public investors, unaware of the hidden go along orders, have no opportunity to adjust their limits in response to such hidden orders so as to maximize their opportunities for a full and complete execution.
Floor brokers thus would be given a huge and unfair competitive advantage over the public limit book order book, which is fully transparent and is the "magnet" which attracts contra side liquidity. Once the contra side liquidity is attracted, however, the go along orders function, in effect, as hidden electronic parasites, leeching off of the public limit orders and, by splitting executions, denying complete executions to the very orders tha attracted the contra side lquidity in the first place.
This is a manifestly absurd and unfair result, and the SEC should not tolerate it. Or maybe the NYSE should just issue the followng disclaimer to the investing public: "Beware. Even though you see the price/time priority of your orders on the public limit order book, hidden orders, that are entered later in time and in reaction to your orders, may well deny you an execution."
THE NYSE HAS NOT EVEN SUBMITTED A RULE GOVERNING AWAY FROM THE MARKET "PARITY"
Shockingly, for all its references to go along orders trading on "parity", the NYSE has not even submitted a rule governing how away from the market go along orders would in fact trade in relation to other orders in the market. The only NYSE rule that describes parity is Rule 72, which describes parity only in the context of the prevailing bid and offer. The concepts in Rule 72 cannot simply be extended to the away from the market scenarios contemplated by the NYSE without a substatial amendment to Rule 72, with opportunity for public comment. (The confusion about away from the market parity is apparent, for instance, in Example 8 in Exhibit 3 of the rule submission. The clean-up sweep price is 20.03, and the orders eligible to split the execution at that price are a go along order to buy 2000 shares at 20.04, and an order on the public limit order book to buy 2000 shares at 20.03. The size of the trade at 20.03 is 3500 shares. Simple parity (which is all the NYSE discus ses) would suggest an even split, one order gettng 1800 shares, and the other 1700 (an exactly even round lot split is not possible here). But in the NYSE's example, the entire go along order of 2000 shares is executed, and only 1500 shares of the public limit order is executed. The NYSE is apparently invoking price priority here, but the literal language of proposed Rule 1000(b)(iii) is vague, imprecise, and does not on its face allow for this result. The result in the NYSE's example appears to be correct, in that it avoids a possble trade-through, but serious redrafting of the rule is in order).
The examples the NYSE gives of how automated go along trading would operate are simplistic in the extreme. The examples provided discuss situations in which only one go along order is involved. Matters become more complicated, however, when, as likely, there are several go along orders on the go along limit order book. One conceptual issue, not addressed at all in the rule submissin, is whether the go along limit order book is considered to be one bidder, or whether each go along order on the go along limit order book is deemed to be making its own, separate bid. In the conventional floor auction, the public limit order book may contain many orders at the same price, but the specialist makes only one bid in the auction on behalf of the book. The NYSE has simply not defined whether (i) the public limit order book continues to be only one bidder, particularly in an away from the market sweep;(ii) each order on the public limit order book is deemed to be a separate bidder; (iii) the go along limit order book is deemed to be only bidder, as the public limit order book is only one bidder; or (iv) each order on the go along limit order book is deemed to be making a separate bid.
The answers to these questions have significant repurcussions as to how orders on both books are actually filled.
Suppose, for example, that the market is .20(bid)--.30(offer), with 1000 shares at the bid and offer. There are four fully transparent orders of 1000 shares each on the public limit order book at .19, and four hidden go along orders on the go along limit order book at .19. An order enters the market to sell 7000 shares at the market.
- 1000 shares would trade against the published bid of .20. If the public limit order book and the go along limit order book are each considered to be making one bid, then each would split the remaining balance of the sell order, with 3000 shares going to each book. The orders on the public limit order book would be allocated 3000 shares in order of time priority, but each order on the go along book would receive an equal split of the 3000 shares.
- 1000 shares would trade against the published bid of .20. If the public limit order book is deemed to be making only one bid, but the each order on the go along limit order book is deemed to be making its own bid, the share allocation picture changes dramatically. In this example, there would then be five bidders, each entitled to receive 1200 shares. The hidden go along orders would be filled completely, and the public limit order book would receive only 2000 shares (as opposed to 3000 shares in the first example).
Clearly, it is in the public interest for the public limit order book, which attracted the contra side liquidity, to be filled before any go along order is executed. If the go along limit order book is permitted to compete directly with the price/time priority of the public limit order book, it should be treated as only one bidder, as there is no rational basis for saying that one repository of orders (the public limit order book) can make only one bid, but another repository of orders (the go along limit order book) can make as many bids as there are orders at the same price. In the event, if the NYSE is to be permitted to allow hidden go along orders to siphon executions away from the fully disclosed public limit order book, the SEC must insist that the NYSE submit a detailed rule governing exactly how away from the market parity would operate. Simply announcing the concept of parity trading is clearly inadequate from a rulemaking perspective when the e xisting parity rule, Rule 72, does not address at all the entirely new context being presented.
THE PROPOSED GO ALONG LIMIT ORDER BOOK IS BLATANTLY ANTI-COMPETITIVE WITH RESPECT TO ORDER ENTRY
Although many large investors are interested in go along/split executions trading strategies, the NYSE is proposing that only floor brokers can enter go along orders onto the go along limit order book. Surely, the only purpose in such a restriction is to protect floor broker jobs and NYSE seat/seat lease prices. Such an anti-competitive position cannot stand.
The NYSE permits qualified institutional customers to enter orders directly into its superdot system if sponsored by a broker-dealer. There is no justifiable reason why such customers should not be permitted to enter go along orders directly onto the go along limit order book, as these orders are not even being executed by floor brokers themselves, but receive automated execution. The NYSE's proposed "5 panel geographical range" makes a mockery of any pretense that a floor broker will be actively managing hidden go along orders on the go along limit order book. The NYSE claims to provide "platforms of choice" to its custmers. Let the NYSE prove it: permit qualified investors (and broker-dealer trading desks) to choose to enter go along orders themselves, directly into the appropriate NYSE system, or to choose to give orders to floor brokers.
Qualified institutional customers (and broker-dealer trading desks) should also be permitted to enter themselves so-called CAP orders (a kind of go along order handled today in formulaic fashion by specialists, but which would also be excuted in automated fashion under the NYSE's proposal).
FLOOR BROKERS AND SPECIALISTS SHOULD NOT BE PERMITTED TO CHARGE FLOOR BROKERAGE ON GO ALONG AND CAP ORDERS THEY DO NOT PERSONALLY EXECUTE
Floor commissions are a senstive subject, and the U.S. SROs are not permitted to engage in rate-fixing. But there is clear, SEC-approved precedent for the proposition that users of automated systems may not charge any commission whatsoever. For example, by SEC-approved rule, NYSE specialists are not permitted to charge floor brokerage on most orders they receive through the superdot system. This prohibition should be extended to CAP orders that are executed automatically, and not by the specialist personally.
Floor brokers similarly should not be permitted to charge floor brokerage on any order (or portion thereof) that is executed by an automated system and not by the broker personally. As the NYSE has drafted its proposal, the following absurd scenario is entirely likely: An institution transmits an order to a floor broker. The broker, busy in another stock in another panel, enters the institution's order onto the go along limit order book. The order is automatically executed, but the broker charges the institution floor brokerage (the institution will probably have no idea as to how the order was actually executed). In essence, the floor broker is charging floor brokerage for performing an administrative task (order entry) that the institution should have been permitted to perform for itself.
THE PROPOSED SPECIALIST DEALER BOOK AND ALGORITHMIC TRADING ARE SERIOUSLY ANTI-COMPETITIVE AND CONFLICT WITH THE AFFIRMATIVE AND NGEATIVE OBLIGATIONS
The ability of specialists to trade in the automated fashion proposed by the NYSE raises serious, fundamental questions as to what the role of the specialist really is (if, in fact, the specialist continues to have much of a role), and whether concepts such as the affirmative and negative obligation continue to have any relevance. The increased reliance on automated trading, and the proposed go along and specialist dealer limit order books, render obsolete the notion that the specialist is really conducting a central auction. The extent to which the specialist continues to be much of an "agent" is highly debatable; the NYSE has simply ignored the question of whether the specialist is the agent for the go along limit order book, although it would appear that he is since he would be representing such orders in what remains of the floor auction when they are part of the displayed bid or offer.
A strking feature of the NYSE's proposal as applied to specialists is the manner in which it resuscitates concerns about being able to trade to proprietary advantage with knowledge not available to other market participants. It is bad enough that floor brokers can enter hidden (from the public) orders that can seriously disadvantage the fully disclosed orders on the public limit order book. But the NYSE is proposing to permit specialists to enter hidden away from the market limit orders onto the specialist dealer book based on his knowledge of both the public limit order book and his unique knowledge of the orders on the go along limit order book as well. For years, institutional investors and others have complained about the ability of floor brokers and specialists to take advantage of information about orders on the public limit order book. After years of fraught discussions, the NYSE developed its open book product, which was largely worthless in its earliest ver sions, but was improved when it began showing the depth of the public limit order book. Now the NYSE is proposing, in essence, to make open book worthless again, because the real depth of the maket in many instances will be on the two hidden limit order books. This is fundamentally a very different situation from that which exists today. While there is obviously information about market depth that is available only on the NYSE floor, such market depth does not take the form of pre-programmed orders that can trade with no degree of public exposure whatsoever, but which are exposed to the specialist, and the specialist alone. And the NYSE would have investors believe this massive march back to the bad old days is actually progress!!
Another striking feature of the NYSE's proposal as applied to specialists is its striking transformation of the dealer function. While the NYSE proffers the expected pabulum about specialists' proprietary transactions having to be effected in accordance with NYSE rules, in fact the proposal tramples on conventional understandings of the affirmative and negative obligations, without any effort whatsover at presenting appropriate rule amendments for public comment.
The ability of specialists to enter orders onto the specialist dealer book, with knowledge of both the public limit order book and the go along limit order book, clearly has the potential to raise specialist penny-jumpng to new heights of creativity, and raises serious questions about their compliance with the negative obligation. In the examples provided in Example 3 of its rules submission, the NYSE speaks of specialists buying on parity with orders on the go along limit order book, but having to yield to orders on the public limit order book. The NYSE is absolutely incorrect here, at least in terms of the unqualified way in which it presented the examples. NYSE Rule 108 (which, as my comment on NYSE-2004-70 demonstrates, the NYSE prefers to ignore) precludes specialists from buying along with anyone (unless covering a short position). The orders on the go along limit order book are as much public orders for purposes of Rule 108 as are the orders on the public limit or der book. The NYSE has not proposed to amend Rule 108. If it were to propose such an amendment, the SEC should look seriously askance. Specialist parity liquidations have been historically encouraged because they facilitate essential re-capitalisation. Specialist parity acquisitions have never been deemed to promote a comparable public purpose, and have not been permitted.
Ultimately, specialist trading by means of the specialist dealer book raises significant surveillance issues, which gives serious pause, given the two major floor trading scandals within the past 10 years. At a minimum, the NYSE order execution facility needs to monitor the specialist's dealer position in real tme to preclude unlawful go along trading activity. Alternatively, if real time position correlation proves impossible, the NYSE can simply require the specialsit's order to yield to all other orders in all instances. But the SEC must demand specific representations from the NYSE in this regard.
While specialist algorithmic trading has a certain appeal (provided that penny jumping is precluded), one aspect of the NYSE's proposal in this regard is a blatant violation of the negative obligation and highly anti-competitive. It is difficult to argue against "price improvement", but what the NYE has proposed would give the specialist a unique trading opportunity, denied to any other market participant, to interact via computer with incoming order flow that no one else in the market would be given an opportunity to react to. Historically, the specialist has been the "price improver" of last resort. The specialist was required to expose the order to the trading crowd to give floor brokers the opportunity to provide price improvement. Only if no one else provided a better price could the specialist do so, ensuring no violation of the negative obligation. Algorithmic price improvement, as proposed by the NYSE, turns the traditional approach on its head and lets the speci alist front run the entire market in this regard, which has never been the specialist's role.
The NYSE needs to open algorithmic price improvement to public competition. Other market participants might well be interested in trading in this regard as well, and more competition can only mean better prices for investors. An NYSE computer could randomly assign the execution when algorithims of more than one market participant would provide price improvement at a given price, with the execution going to the specialist only when no other algorithm would provide price improvement at that price. In the event, the NYSE needs to propose for public comment a detailed, specific amendment to the negative obligation if the specialist, and only the specialist, is to be granted this exclusive proprietay trading opportunity.
The NYSE's proposal also raises serious questions regarding the specialist's affirmative obligation to trade as necessary to maintain a fair and orderly market. In the current floor auction, the specialist will typically be actively engaged in the fair pricing of orders received via superdot, on one side of the market, whose size exceeds that of the published bid or offer. The specialist will assess the depth of the public limit order book, and the extent of possible trading crown participation and attempt to set a price for the order that is appropriate in relation to the size of the remaining balance of the order, the last sale price of the subject security, and its typical trading characteristics. The specialist will, as necessary under the affirmative obligation, commit capital in the price setting process. When there is an away from the market clean-up price, the specialist is required under the affirmative obligation to commit capital to ensure an orderly aftermarket.
The NYSE's proposal is shocking in its absence of any discussion whatsoever as to how the specialist would discharge his affirmative obligation when the remaining balance of an order (which may not even be that large) is swept to an away from the market clean-up price that is not reflective of a fair and orderly market. The notion that the specialist "may" have pre-programmed orders on the (hidden) specialist dealer book is far too vague to give any comfort that markets will, in fact, reflect the NYSE's traditional concerns about price continuity in relation to overall market depth. And the NYSE's pro forma representations about specialists having to conduct business in accordance with applicable regulations lack the specificity required from a rulemaking perspective, particularly given the increasing lack of public respect for the NYSE's somnolent surveillance department.
In sum, a serious deficiency of the NYSE's rules submission is that it simply announces concepts regarding the specialist's ability to enter pre-programmed orders and trade algorithmically but does not present a rule squaring those concepts with the specific requirements and constraints of the affirmative and negative obligations. The SEC simply must demand an additional submission from the NYSE in this regard, to be published for public comment.
THE PROPOSED PRICE IMPROVEMENT METHODOLOGY CONFLICTS WITH EXISTING RULES, WILL RESULT IN EXECUTIONS THAT ARE INFERIOR TO THOSE AVAILABLE TODAY IN THE FLOOR AUCTION, AND WILL PERMIT THE SPECIALIST TO REPRESENT AN ILLEGAL 'NOT HELD' ORDER
The NYSE's proposed price improvement methodology for market orders and improved price limit orders (I refuse to use their counter-intuitive term "auction limit order") will clearly result in executions (if the order is in fact executed) at prices significantly inferior to what those orders would receive in the floor auction, and deprive those orders of the protection they have currently of being represented by an agent who is "held" to the market when the order is presented for execution.
The NYSE is proposing a method of exposing orders for price improvement that conflicts with the crossing and order exposure requirements of NYSE Rule 76 (specialist is agent for orders on both sides of the market) and NYSE Rule 91 (specialist is agent for one order, and would execute that order against the dealer account). Under both rules, the specialist is required to make a bid that is one increment below the offer (or an offer one increment above the bid) before completing the trade. Intutitively, this method of order exposure is most likely to result in price improvement, which is, presumably, why orders have been exposed for price improvement in this manner for eons. (Discussions I have had with NYSE staff confirmed my intuitive suspicion). Under the NYSE's proposal, however, a bid would be made one increment above the existing bid (as opposed to one increment below the existing offer, as under Rules 76 and 91). Without any discussion whatsoever, much less appropriate amendments to Rules 76 and 91, the NYSE is proposing a new method of order exposure which lessens the likelihood of price improvement, but increases greatly the prospect that orders seeking price improvement will in fact be executed at prices far worse than those existing in the market when the order was first presented for execution.
A simple example shows the contrast. Suppose the market is .20(bid)--.30(offer), with 1000 shares at the bid and offer, and the specialist receives a market order (or a marketable limit order) to buy 1000 shares. Under current rules, the specialist is required to announce a bid of .29, and, if no one traded there, would complete the trade at .30. The process might take a second or two. Under the NYSE's proposal, however, the incoming buy order would be bid at .21 for as long as 15 seconds, and later-arriving orders could exhaust the .30 offer, resulting in the buy order being executed at a price above the .30 offer, "the market" existing when the order first arrived. While in theory the order might receive a greater degree of price improvement under the new proposal, the NYSE's own history of Rule 76 and Rule 91shows that price improvement is more likely when the bid is more closely related to the offer, as currrently required.
The NYSE's own examples demonstrate the absurdity of the NYSE's approach. My personal favorite is Example 5, in which a marketable limit order to buy, with a limit price of .12, is immediately executable against a contra side offer of .09. Under current rules, the specialist would announce a bid of .08, and the trade would take place immediately at .08 or .09. Under the NYSE's proposal, however, the marketable limit order is bid at .06 (one penny better than the .05 bid pre-existing in the NYSE's example). The marketable limit order is bid for as long as 15 seconds. In the NYSE's example, an auto ex order to buy systemically exhausts the .09 offer, and all the liquidity on the book, during the 15 second order exposure period for the earlier-arriving marketable limit order. After the auto ex order exhausts the liquidity against which the marketable limit order could have been executed, the marketable limit order becomes the published bid at its limit price of . 12. If the market continues to move up, the order would not be executed, even though it was immediately executable, at a price better than its limit, when it was presented for execution. And the NYSE presents this without embarrassment, as though somehow it were rational!!
No marketplace that I'm aware of treats orders in so ridiculous a fashion. The NYSE's proposed "price improvement" methodology is actually a prescription for worse prices, as the NYSE is effectively redefining "market" as, not the price that people see on the screen, but the price that may exist 15 seconds after the order is presented for execution. I cannot imagine that broker-dealers, who have serious "best execution" responsibilities under SEC rules, will want to play "Russian roulette" with their customer order flow in this fashion (maybe a better price, a good possibility of a worse price).
Beyond the sheer incoherence and unprecedented nature of the NYSE's proposal, there are also readily apparent legal problems. The NYSE takes the common sense position that the specialist is not the agent for pure auto ex orders because these orders do not enter the floor auction. However, the specialist is the agent for any order that is actually systemically delivered to the specialist's trading computer on the trading floor. The specialist must be the agent for the price improvement orders in its proposal because NYSE rules require an actual member as the agent for each side of a trade executed in the auction (as the price improvement orders could be).
Under both SEC and NYSE rules, the specialist is considered to be a "held" agent, meaning that an executable order represented by the specialist is entitled to no worse than the published contra side price when the order is presented for execution. There is a common sense exception if there are earlier- arriving orders in the market that have a prior claim on the contra side liquidity. If the specialist does not execute the order promptly, and later-arriving market interest exhausts the contra side liquidity, the specialist is "held" to giving the customer the price the customer should have received had the specialist acted promptly. In the example above, if the specialist "missed the market" under current rules when a later-arriving order exhausted the contra side liquidity, the specialist would be required to make the improved price customer whole at a price of .09.The NYSE acknowledges in its rule submission that price improvement orders may "miss the market", bu t fails to acknowledge that it is proposing thereby to give specialists an illegal "not held" order to the detriment of public investors.
The prohibition against specialists representing "not held" orders promotes two significant public purposes.
- Investors entering orders take comfort that the specialist, as a "held" agent, will get them the best price in the market when an order is presented for execution, or else will make them whole at the best market price if he fails to do so.
- The prohibition against specialists representing "not held" orders addresses a potential conflict of interest that may arise if a specialist, for example, fails to promptly execute a buy market order against the current offer, lets a floor broker take the offer, and then sells to the buy order from the dealer account at a higher price. This conflict of interest continues to be present in the NYSE's proposal. While specialists cannot place orders on specialist dealer book in reaction to a specific order, they clearly would be permitted to pre-program algorithms that "re-quote" the market at inferior prices when later-arriving orders exhaust liquidity to the detriment of price improvement orders during the 15 second window. In the event, a clear, simple customer protection is being lost here.
Regardless what the NYSE thinks it is up to, the SEC cannot permit this approach as a matter of law, and I cannot imagine that either the buy side or the sell side will stand for this as a matter of proper order execution and discharge of specialist fiduciary duty under the law of agency.
THE PROPOSED SWEEP METHODOLOGY WILL RESULT IN EXECUTIONS LESS FAVORABLE TO INVESTORS THAN THOSE WHICH EXIST IN THE CURRENT FLOOR AUCTION
In the current floor auction, a floor broker with a large order can, after trading with the published bid or offer, follow one of two alternative approaches regarding the remaining balance of the order:
- The broker may seek one clean-up price for the remaining balance.
- The broker may trade at various intervening price levels (depending on the liquidity on the public limit order book or trading interest of other floor brokers) to obtain a better overall price for the order, rather than one clean-up price, which is by definition the worst available price.
With no discussion whatsoever, the NYSE is simply proposing to eliminate the second alternative for anyone needing the speed of automated execution. Such customers would be, in essence, forced to accept the worst price. If the NYSE is truly serious about offering "platforms of choice" to its customers, it should permit the same range of choices in its "hybrid market" that it permits in its floor auction. Customers should be permitted to choose to receive a better overall price by sweeping the book at intervening price levels, or sweeping the book at one overall clean-up price.
There is no disadvantage to orders on the public limit order book if a customer trades at intervening price levels because those orders have declared their intention to trade at their limit prices. Indeed, the NYSE itself has historically seen no disadvantage to such limit orders, as it has long permitted floor brokers to trade at intervening price levels against the public limit order book rather than trade at a single clean-up price.
In the event, it makes no sense whatsoever for the NYSE to permit one type of methodology in its floor auction, and another in its electronic system. At a minimum, the NYSE needs to rationalise the disparate treatments. Far more preferable would be a re-thinking here to give electronic customers the same set of choices they have in the floor auction.
The NYSE's proposal is materially deficient and incomplete. The SEC must insist upon the following:
- Public limit orders on the public limit order book that are fully transparent and disclosed should have priority over hidden go along orders.
- Qualified customers should be permitted to enter go along orders directly into NYSE systems.
- Specialists and floor brokers should be precluded from charging floor brokerage for orders they do not personally execute.
- The NYSE must submit rules governing away from the market parity and how it is modifying the affirmative and negative obligations.
- The NYSE , as the primary market in its listed stocks, must be made to provide price improvement methodolgies that are at least as favorable to investors as those which currently exist.
- The specialist must be "held" to the market at all times when representing market and marketable limit orders for price improvement.
- Algorithmic price improvement trading opportunities must be made available for all market participants, not just the specialist.
- Customers must be given the choice to sweep the book to receive the best overall prices for their orders.
I would appreciate it if the SEC staff would forward to me by email any response the NYSE may make to the above.
Very truly yours,
Do you Yahoo!?
Yahoo! Mail - now with 250MB free storage. Learn more.