February 18, 2005
The author has been consultant to several non-U.S. securities markets in their transition from floor-based to screen-based trading. The author is writing on behalf of two institutional trading 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 SEC and U.S. SRO regulations, as well as several NYSE specialists and floor brokers.
The rule submission at issue, while seemingly technical and obscure, actually represents something quite dramatic, although one would not immediately realise this from the way the NYSE has disguised the submissions true import. For apparently the first time, the NYSE is taking the position that when the limitations of its technology collide with the practical demands of its floor auction, such floor auction and the NYSEs much-hyped price improvement model will be readily discarded.
As shocking as the departure from the price improvement model is, the rule submission additionally attempts to codify a truly bizarre notion. The NYSE is proposing that an order must participate in a trade even though the order was not even in the marketplace when the trade took place, and thus could not have, under any circumstances, participated, or been entitled to participate, in the execution process, or been a party to the contract to trade. Such a concept is simply unknown in any securities market, anywhere.
In the highly unlikely event it actually approves this mess, the SEC must carefully rationalise the radical new law it would be making.
At the outset, before discussing the specific rule submission at issue, I wish to expres my dismay at a disheartening tendency in recent NYSE rule submissions to the SEC. The NYSE seems intent on promulgating new rules without regard for any of its existing rules that may be in conflict with a new proposal. The NYSE does not even identify the conflicting rules, perhaps to avoid the unpleasant task of dealing with their background, history, and purpose, the consideration of which might well render the new proposal somewhat, or even highly, suspect. One deleterious result is obvious - a rule book littered with conflicting rules, the nightmare of every compliance officer. Another result is insidious. By failing to even mention much less amend a conflicting rule, the NYSE is effectively deceiving the public, which presumably lacks expertise in NYSE rules, and which is thus denied the opportunity to respond intelligently when the SEC publishes a proposal for comment. The SEC staff itself may be kept somewhat in the dark as well when the NYSE thus fails to provide a complete perspective as to every rule impacted by a proposal, and why a proposed amendment is justified,notwithstanding the purpose of a conflicting rule. Furthermore, the NYSEs actions appear to violate the SECs own rules, which require an SRO to identify all rules at issue with respect to a particular rule amendment submission.
The NYSEs disregard of the SECs requirements was quite openly announced in Amendment No. 1 to SR-NYSE-2004-05, in which the NYSE made the statement that the proposal being submitted therein would supersede any of its inconsistent rules. This sort of position simply will not do. The SEC must insist that an SRO identify all rules that may be in conflict with, or inconsistent with, a rule amendment submission, and appropriately discuss and amend those rules. SR-NYSE-2004-05, and its two amendments, the hybrid market proposal, are suffused with ludicrous jargon, are treasure troves of conflicting rules, and will be the subject of a separate comment.
The NYSEs position, while not announced in other rule submissions, is an underlying theme, for example, in SR-NYSE-2004-70, the subject of this comment. SR-NYSE-2004-06 provides a particularly blatant example of the NYSEs misleading rule amendment process, as discussed immediately below.
In this rule submission, the NYSE proposed to adopt an amendment to Rule 104 to provide that, when liquidating or decreasing a position, specialists may not trade on parity with a floor broker representing a customers order in the trading crowd if the customer refuses to permit the specialist to do so. Parity is the NYSEs euphemism for the ability of floor brokers to trade side-by-side with each other and the public limit order book and thereby split trades, regardless of the fact that orders on the public limit order book were entered earlier in time. Most investors are unaware that their orders on the public limit order book are not executed in strict time priority with respect to all other trading interest, as they would be in most other markets. When there are no orders on the public limit order book, the specialist may trade side-by-side with floor brokers if liquidating or decreasing a position.
In its rule amendment submission, the NYSE stated that specialists have long been permitted to engage in such side-by-side trading. Quite surprisingly, however, the NYSE neglected to mention, much less amend, the prominent NYSE rule that gives specialists this permission, Rule 108. That Rule 108 clearly applies cannot be in doubt both SEC and NYSE staff had confirmed this point to me a number of years ago in connection with a project I was then working on.
NYSE Rule 108 applies to all members permitted to initiate orders on the floor, which is the NYSEs traditional jargon for various forms of on-floor trading. The most prominent
NYSE on-floor traders, of course, are specialists, who initiate orders when they effect dealer trades. Most orders initiated on the NYSE floor are initiated by specialists, but the rule also applies to the NYSEs Competitive Traders, Registered Competitive Market Makers, and floor brokers covering an error or engaging in arbitrage.
By not even mentioning Rule 108 in its rule amendment submission, the NYSE avoided having to discuss the rules background, history, and purpose. Furthermore, the NYSE avoided having to discuss the potentially anti-competitive nature of its proposal the proposals applicability only to specialists, and not the other on-floor traders covered by the rule.
Presumably, the over-worked SEC, in approving the proposal, relied on the NYSEs good faith in representing that no other rules were impacted by its proposed amendment to Rule 104, and regrettably failed to focus on Rule 108.
But the onus here is clearly on the NYSE. With the SECs approval of the NYSEs proposal, the NYSE now has a rule book that says that specialists can engage in parity liquidations without restriction Rule 108 and another rule Rule 104 that in fact imposes possible restrictions.
The purpose of the parity liquidation provision in Rule 108, as applied to specialists, is simple and to the point: it is intended to facilitate specialists re-capitalisation so that they have the financial resources to continue to be able to make markets, The rule is of long duration. One SEC staff member, a number of years ago, directed my attention to the SECs 1963 Special Study of the Securities Markets, in which the SEC noted, perhaps only half-jocularly, that Rule 108 was consistent with the specialists divine right of liquidation. In the event, the SEC over the years has apparently been quite comfortable with specialist parity liquidations under Rule 108. This makes sense for two reasons:
1. Restraints on market maker liquidations are scrutinised closely by regulators world-wide, as it is self-defeating in terms of overall market quality to artificially inhibit re-captialisation.
2. On the NYSE, so-called parity transactions are typically effected at last sale prices, so specialists parity trading does not directly influence price movements. Under the NYSEs amendment, however, specialists who cannot trade side-by-side with other trading intrest at last sale prices will have to initiate price changes to liquidate positions, which both the SEC and the NYSE have historically attempted to minimise.
One would have thought, as a matter of honour and of compliance with SEC regulations, that the NYSE would have proposed to amend Rule 108 and made a case that its proposed restraint on specialist recaptilisation is somehow justified today, even though the absence of such a restraint has been deemed appropriate for eons in terms of enhancing specialists market making ability.
The NYSEs only rationalisation for its proposal was a statement about public orders trading directly with each other. This position is clearly inadequate such a claim could obviously have been made when Rule 108 was adopted and at any tme since, but the prevailing view clearly has been that a greater public good is served by efficient specialist re-capitalisation.
Ultimately, of course, this is a matter for the SEC to resolve. My point is simply to object to the misleading manner in which the NYSE made an end run around a long-standing rule.
As the NYSE has failed to act honorourably in this matter, the SEC should consider the issue de novo, insist upon a properly presented and argued NYSE rule submission, and publish the proposal for comment so that the investing public may fairly evaluate the effect of the proposed restraint on specialist re-capitalisation on overall NYSE market quality.
This rules submission is hugely significant, but in its typical recent fashion, the NYSE has failed to even mention the key rule involved, much less present in context the rules background, history, and purpose.
On its face, the NYSEs rule submission appears simple. If a specialist effects a dealer trade in the floor auction, but before it is reported to the tape the specialist receives a superdot order executable at that price, the specialist must yield the execution to the subsequent superdot order.
The proposed amendment to Rule 104 mandates this yielding in absolute terms strictly required, irrespective of the actual market dynamic when the subsequent superdot order actually arrives on the trading floor. The proposal provides that ...dealer transactions by a specialist that have not yet been reported by the specialist must yield to any order or orders received through an Exchange order delivery system after the oral commitment to transact, provided that such order or orders are capable of trading in place of the specialist in the consummated transaction. According to the NYSE, if the subsequent superdot order can participate, it must participate, no exceptions.
The genesis of the NYSEs proposal apparently is some sort of technological limitation in the software used by specialists in the computer in which they receive orders and process trades display book in NYSE jargon. It appears from the rule submission which is not very clear on any of this that technical software problems either make it difficult for specialists to report trades when they subsequently receive executable superdot orders, or confuse the NYSEs surveillance systems as to whether the specialists trades were proper in the first place.
Whatever the problem really is, the NYSE has proposed the worst possible resolution from a public interest standpoint, for the following reasons:
1. The notion that an order that was not even in the marketplace when a trade was completed must nonetheless particpate in the trade is bizarre, unprecedented, completely at odds with how marketplaces really work, and of highly dubious legal validity.One shudders to think of where such a counter-intuitive precedent could lead.
2. The proposal does not even distinguish between superdot market and limit orders, which, as discussed below, need to be treated quite differently under the NYSEs own rules.
3. The proposal fails to even mention, much less amend, its key auction market rule, Rule 76, which mandates that orders be treated differently from what the NYSE has proposed. NYSE Rule 76 as has been explained to me repeatedly over the years by NYSE specialists and floor brokersrequires crosssing of orders when an agent is representing both a buy order and a sell order. The rule clearly applies to a specialist, for example, when the specialist is representing a superdot limit order published as the NYSEs bid or offer, and a superdot market order that would trade against it. The purpose of Rule 76 is to seek price improvement for an order. The NYSE prides itself on price improvement its marketing propaganda consistently emphasizes that a high percentage of superdot market orders receive better prices than the bid or offer existing in the market when they arrive for execution. The NYSE rule submission effectively would dispense with Rule 76s crosing/price improvement procedure, and assign a price to subsequent superdot market orders without giving them an opportunity to receive a better price. And all of this is proposed to be done without even a single mention of Rule 76s requirements
4. By not even mentioning Rule 76, the NYSE has conveniently avoided having to acknowledge that its proposal may cost investors millions of dollars, much less justify that likely result.
5. The NYSE has misused the term yield. What the NYSE means is a substitution of principals.
In order to understand what the NYSE is proposing, it is important to consider the different ways that a specialist represents market and limit orders. If the quotation is .50bid--.60offer, and the specialist receives a superdot limit order to sell with a limit price of .58, NYSE autoquote will change the quotation to .50bid--.58offer. However, the specialist may improve the market by making an offer of, say, .56. If a trade takes place at that price, all well and good. The specialist has acted under colour of right. The NYSE is apparently concerned, however, with a situation in which the specialist, under colour of right, has lawfully traded at .56 but, before the trade is reported to the tape, he or she receives a limit order at .56. The NYSE is proposing that even though the specialists trade was entirely legal and in fact may have been in furtherance of essential re-capitalisation the specialist must yield the execution to the subsequent superdot limt order. This is a radical departure from the traditional first come, first served auction that has otherwise served the NYSE in good stead for more than 200 years, and which is the way all markets operate. Surely assuming the SEC takes this nonsense seriously in the first place the SEC will demand greater justification from the NYSE for such an historic sea change than technical software problems.
Aside from the conceptual muddle, there are other problems with what the NYSE has drafted.
1. The NYSE has misued the term yield, which is sure to create confusion. Traditionally, when one speaks of yielding, one means to defer to other market interest which will then actually consummate the trade that the yielder has had to foreswear. In other words, the yielder steps aside, does not trade, but lets another market participant consummate the trade. This is the traditional meaning of yielding. See, e.g., Section 11a1G of the Securities Exchange Act. The NYSE rule submission presents the situation in which the specialist has already consummated a trade. The subsequent superdot order does not consummate the trade, but rather would simply be substituted as the contra party. There is no yielding in the trading process itself, as that term has always been used. The NYSE needs to make clear that it is proposing a substitution of principals in a consummated transaction, not a yielding.
2. The NYSE has failed to discuss the size of the specialists trade in relation to the size of the subsequent superdot order., and the potential for havoc in the trade settlement process.
Suppose, for example, that the specialist has sold 1000 shares to floor broker A, who then leaves the trading area. Before the trade is reported to the tape, the specialist receives a limit order to sell 500 shares at the price of the unreported trade. The NYSEs proposal would appear to require the specialist to substitute the superdot limit order as contra party to the trade to the extent of 500 shares, with the specialist being the contra party for the other 500 shares. Or does the specialist report two trades to the tape unlikely? Either way, broker A will be seeking to settle the trade only with the specialist, not a second contra party the subsequnet superdot order. This is clearly a step backwards in terms of oerall market efficiency.
In its application to market orders, the NYSEs proposal is entirely inconsistent with the underlying philosophy of its floor auction and the specific requirements of Rule 76. A few simple examples many more could be provided should suffice to drive the point home.
1. Suppose the quotation is .50bid--.60offer, with the bid size being 1000 shares. The specialist sells 1000 shares to the bid at .50, but before the trade is reported to the tape the specialist receives a superdot order to sell 1000 shares at the market. Clearly, this order is capable of participating in the transaction which the specialist just consummated, which is how the NYSE drafted its proposal. Just as the specialist sees the subsequent superdot market order to sell, however, NYSE autoquote publishes a new bid of .55 for 1000 shares. Under the NYSEs proposal as drafted, the specialist would be strictly required to substitute the subsequent superdot sell market order as the contra party to the trade at .50, even thoughthe market for that order had become .55, an improved price for the order. The specialist having substituted himself or herself out of the trade at .50could then trade with the new bid of .55, grabbing the improved price that should have gone to the subsequent superdot order. It is difficult to imagine that the NYSE could have intended so ridiculous a result, but the simplistic, absolutist drafting of its proposal mandates this result.
2. Suppose the market is .50bid--.60offer, with the size of the bid being 2000 shares of orders on the public limit order book for which the specialist is the agent. The specialist sells 1000 shares to the bid, leaving a bid of .50 for 1000 shares. Befoe the trade can be reported to the tape, the specialist receives a superdot market order to sell 1000 shares. Under the NYSEs proposal, the specialist must simply substitute the subsequent superdot market order as the contra party in the trade at .50. However, Rule 76, one of the pillars of the floor auction, requires the specialist to make an offer of .51 on behlf of the subsequent superdot order to try to obtain price improvement for the order. As drafted, the NYSEs proposal does not allow for the possibility of price improvement, but mandates that the specialist simply assign the price of .50 to the subsequent superdot order.
Both examples above show simple instances where a superdot order arriving after a specialist has consummated a trade can in fact suffer economic harm under the NYSEs proposal. Since the NYSEs own statistics show almost half of superdot market orders receive price improvement, it is clear that the economic harm from the NYSEs proposal could potentially run into the millions of dollars when aggregated.
More skilled, precise, and comprehensive drafting could, perhaps, result in a somewhat better case being made for what the NYSE is proposing. But not even the most polished draftsmanship can render respectable the absurd concept at the heart of the NYSEs proposal, namely that orders not even in the marketplace when a trade is consummated get to participate in that trade anyway.
This notion has obviously been formulated in response to technological deficiencies and surveillance inadequacies. That such tail wagging the dog factors could undermine fundamental auction market principles gives serious pause to that ever-dwindling minority that would like to believe, despite all contemporary evidence to the contrary, that the NYSE knows what it is doing.