May 5, 2000

Mr. Jonathan G. Katz
Securities and Exchange Commission
450 Fifth Street,
N.W. Washington, DC 20549-0609

(Release No. 34-42450; File No. SR-NYSE-99-48)

Self-Regulatory Organizations; Notice of Filing of Proposed Rule Change by the New York Stock Exchange, Inc. to Rescind Exchange Rule 390; Commission Request for Comment on Issues Relating to market Fragmentation.

Dear Sirs:

Scudder Kemper Investments leads the global investment management business of the Zurich Financial Group, one of the world's foremost providers of risk management, asset management, and other financial services.

With more than $280 billion currently under management and almost 80 years of experience Scudder Kemper Investments is among the world's largest and most experienced asset managers. Scudder Kemper's client base includes institutions, individual investors and financial intermediaries worldwide. We trade equities in sixty-nine different markets around the globe and staff a 22-hour trading desk out of our offices in New York City.

Scudder Kemper Investments welcomes the opportunity to comment on the SEC's "Self Regulatory Organizations; Notice of Filing of proposed Rule Change by the New York Stock Exchange, Inc. to rescind Exchange Rule 390; Commission request for Comment on Issues Relating to Market Fragmentation."

Today we are seeing the current market model of a central trade destination evolve toward a more decentralized structure. The ultimate result of this evolution which we feel would best serve all types of market participants is a virtual limit order book with price/time priority, trade-through and trade-at rules. Scudder Kemper Investments supports this evolution toward open trading architecture, but we realize that competition must also exist among market technologies and venues to ensure the most effective and efficient market outcomes. Although we cannot fully support any one of the SEC's options for addressing fragmentation in the concept release letter, we do support the following ideas:

Limit orders need to be protected and encouraged because they provide the pricing foundation in the equities market. The establishment of clear and effective trade-through and trade-at rules to protect price/time priority in each market center would be required to accomplish this goal.

In the linked market environment that we envision and support, we agree with Chairman Levitt's view that Rule 390 should not be part of the future of the securities markets. The most difficult issue in this new market model will be maximization of order flow convergence. An appropriate balance needs to be struck between entrepreneurs and regulators. New businesses will create technology and capitalize on market niches; regulators must harness this creativity to achieve maximum trade flow convergence and the resulting liquidity.


The markets are already dealing with fragmentation through technology. The FIX protocol1 is a good first effort toward electronic interaction. Brokerage firm traders use this protocol to advertise indications of interest to their customers, the institutional traders at investment management firms. In addition, FIX offers a uniform language for order transmission. Both institutional and retail brokers support the use of this protocol.

Order management facilities have allowed for traders to move large flows of orders in and out of ECN's and crossing networks, as well as transmit and receive orders through individual brokers. Some institutional desks have built their own order management systems; others such as Scudder Kemper, Investments, have used vendors that supply sophisticated solutions. Some of these order management systems allow for straight-through processing, including routing orders to a broker of choice.

ECN's have created linked features to other electronic brokers and exchanges. Archipelago has created the technology to seek best price in the marketplace. Another capability created by the ECN's is a sweep feature allowing the trader to interact with the entire electronic market simultaneously. Reserve features have allowed for the institutions to keep their anonymity. The internet has allowed for the retail investor to get closer to the market at a lower cost resulting in institutional and retail order flow becoming interdependent.

In response to the SEC's concern regarding fragmentation, we believe that the progress made to date, as cited above, has been positive. We feel that the forces of creative competition will fuel a continued evolution toward an efficient, technologically linked market. The regulatory challenge will be to recognize that a virtual market is fundamentally different than our current structures. Regulation will need to foster continued technological advances.

A virtual, global limit order book, has no time boundaries. After hours trading, if demanded, could further fragment the market. To date the increased evening trading hours to satisfy the retail investor has not created the demand for further staffing of the Scudder Kemper trading desk. We do have two concerns about this form of fragmentation:

Scudder Kemper Investments, would encourage this virtual market linkage to expand hours in the morning to better overlap with Europe. Rather than an evening trading expansion, a morning expansion would better capture the important European trading timeframe. The U.S. primary exchanges and NASDAQ should take a global leadership position during these hours.

Rule 390

The elimination of Rule 390 is logical given the market evolution we are experiencing. Rules that mandate location-specific trading lose relevance in a virtual market. With the elimination of Rule 390, strict price/time priority and open architecture in each venue, at a minimum, need to be established to protect exposed limit orders. Internalization is the direct result of today's inadequate lack of support for price/time priority. Our goal should be to maximize order confluence to achieve best execution for the customer, not to maximize profitability for the agent. Limit orders in the core market are the backbone of the virtual limit order book. Currently these orders are treated as options for all to deal around. The value of these orders is not recognized by the current functionality of the market. Weakness exists in the current linkages of the various competitive trading venues. Each member of the NLS should have the same rules governing price/time priority in their system as well as responsibility to connect to the best price in the NLS. We support Rule 390's elimination. We believe that there is an opportunity through proper regulation to better deal with limit orders and allow for order confluence.

A two-tiered structure of regulation would be desirable. The SRO element would be responsible for policing its own system in addition to their current regulatory efforts. An oversight regulatory body, made up of a fair representation of the investing community, should set minimum standards for the technologies responsible for linking and enforcing order interaction.

Competition should determine the best linkages. Those that do not meet the minimum technology standards should be encouraged to upgrade. These linkages are the delivery mechanism of best execution. The regulators should require and oversee best execution. This regulatory attention to the functioning of the market should promote the proliferation of the most efficient and effective linkages.

ITS is an example of a structure that offers inadequate linkage. ITS needs an independent regulator to correct its inadequacies. ITS, which was designed to link existing markets, does so inadequately because of its rules and its governance. Total unanimity is required to enact any proposed change. This set-up discourages change that would be best for the greater good at the expense of the individual exchange. With this form of governance in place, evolution is at risk.2

An antiquated rule exists in ITS which gives the specialist 60 seconds to respond to an ITS order. In an electronic world or high-volume marketplace, 60 seconds is an eternity. This rule clearly favors the NYSE specialists and allows them to deliberately slow the trading process, to their advantage. To change this rule would be very difficult in the current environment. Complete and efficient market linkage is critical to make electronic commitments across the marketplace work. If we move to an electronically integrated market structure, rules would need to be altered. An initial hurdle would be to change the ITS governance.


Scudder Kemper Investments supports regulation that encourages order interaction for all market participants. Internalization assumes the price, without taking into account quantity, on posted limit orders is the market price. Trading on these prices, even in small quantities can change the quote. Internalization becomes profitable for the dealers at the expense of the core market. If broker dealers were crossing stock at a level that represented price improvement for both sides, internalization would not be a problem. In practice, there is potential for the broker dealer to gain price benefits that, because the transaction occurs outside an active market, are not available to the client.3

The NASDAQ market openings are a clear example of the problem caused by internalization. The lack of a structured opening process causes price disparities and volatility due to internalization of orders. There is a lack of public interaction of orders at the NASDAQ open and close. Single priced openings and closings in NASDAQ are necessary and could help to avoid this problem. Without moving these orders to a single priced auction, large orders do not get to interact with the small retail orders and mis-pricing occurs. The run-off of matched orders in the NASDAQ market on the opening confuses our portfolio managers and traders. The inability to interact is frustrating and may effect our ability to offer best execution to our clients.

The retail investor, due to advances in technology, is offered efficient and inexpensive systems to trade in the marketplace. Payment for other non-execution services should not be accomplished through execution inefficiencies, but rather through specific fees. The negative impact of veiling non-execution costs in the execution outcome is felt not only by the individual investor, but by the general market as well. The supply and demand in the market place should always dictate the price of a transaction.

Scudder Kemper Investments questions how an agent who is responsible for providing best execution can in fact ensure this result when the actual trade is sold to another broker. Are they selling the order flow for the best deal or for who affords best execution? Non-dominant market centers that cannot afford the proper linkages to the liquidity pools and compete on a service basis should continue to be allowed to farm out their execution. Disclosure of best execution should be required.


Decimalization should decrease the value of payment for order flow. Decimalization will narrow spreads on the most liquid stocks. As the spread is reduced, the interest in this flow will also diminish because of declining potential profit.

Greater market transparency is a requirement as decimalization is put into place. The exposure of the NYSE specialist book and NASDAQ super montage must be complete. The NYSE's Institutional Xpress is a good first step in this long-overdue process. Along with exposure of the specialist book, the NYSE Superdot system needs to become an automatic execution vehicle. The floor, under the current conditions, does not offer a level playing field. The floor broker has standing over the electronic delivery system. Electronic orders are consistently "stepped in front of" as the specialist tries to improve price, when the customer would have been satisfied with the exposed price.

If the NYSE moves to time/price priority, rules such as the clean-cross rule would need to be eliminated.4

Conclusion Scudder Kemper Investments supports competition amongst market venues that encourage:

The problem of fragmentation in the current market is being addressed through advances in technology. The concept of one central limit order book is being replaced by a virtual limit order structure. The various trading venues and exchanges need minimum linkage requirements in order to allow the traders to offer their clients best execution.

Scudder Kemper Investments supports the New York Stock Exchange elimination of Rule 390. Securities traded on the New York Stock Exchange need competition from other exchanges and electronic networks to ensure that an efficient structure exists.

We believe that strict price/time priority and the support for limit order protection should be enforced within each market structure. Limit orders are important and should receive priority in each venue. Complete transparency is essential for all participants both institutional and retail in order to support this structure.

Internalization without price improvement should be eliminated. Decimalization should encourage those that need price improvement to transact orders. The minimal spread caused by decimalization should help to eliminate the payment for order flow.

The possible options offered by the Securities Exchange Commission to address fragmentation each offer partial solutions. I hope that our recommendations and comments offer a clear idea of what an institution needs to enhance its ability to offer best execution to its clients.


Peter W. Jenkins

Managing Director


1 The FIX protocol was created by a committee comprised of members of both the brokerage and the investment communities. It is a common language for system interaction.

2 Optimark is an example of an electronic innovation that may fail, partially because of the antiquated rules of linkage and the governance at ITS.

3 The structure of payment for order flow leads to internalization that borders on a conflict of interest, particularly at the open. The relationship between retail and wholesale brokers has become overly dependent to the exclusion of the open market. Please see WSJ article by Greg Ip, "Forget Spreads, This Is Better," 3/3/2000.

4 Another form of internalization.