We use the term "market data business practices" to refer to the whole spectrum of issues associated with market data policy, contracts, billing/reporting and administration. The two core questions before the Advisory Committee are (1) who should determine the business practices of the exchange and (2) to what degree should there be regulatory oversight over those business practices?

Some of the more important business considerations below may help the Advisory Committee better evaluate these questions:

Prior-Approval versus Vendor-Discretion: In a very broad sense there are two types of business models being used by global exchanges. The first (vendor discretion) gives a license to the vendor to redistribute data without pre-approval but subject to the terms of the exchange contract. This is a model used by most exchanges around the world. The second (prior-approval) starts with the premise that no one is allowed to do anything with the data until it is approved in advance by the exchange. This is a model used primarily by the North American equity exchanges. Both business models are subject to audit as the means of verification. Some vendors have argued that the North American equity exchange approach is administratively cumbersome. The exchanges contend that it is necessary for them to effectively manage their business.

User Classification: North American exchanges use various classification systems to determine the rights to use market data and fees. For example, there is a distinction between "professional" and "non-professional" subscribers for fee determination. In addition, there are distinctions between "vendors", "sub-vendors" and "subscribers" as well as distinctions between "internal redistribution", "external redistribution" and "no redistribution". These classifications are used to determine contractual obligations, liability, fees and administrative requirements. For each class of user, fees may vary based on how the user applies the data.

Unit of Count: This refers to how market data is counted and priced as well as to how the unit is to be applied in vendor and exchange contracts. There is an ongoing debate within the industry on who and what gets counted. Units of count can be device-based, user-based, location-based or identification-based. There are units of count policies for single users, common areas, shared terminals, and users with multiple terminals. The unit of count is not consistent within the industry. Implementation is dependent on billing approach (direct versus indirect) and exchange business policy. Inconsistent units of count definitions complicate billing reconciliation.

Subscriber Agreements: Subscriber agreements are the method by which exchanges and vendors seek to control the use and distribution of information and determine responsibility for compliance with exchange requirements. Subscriber agreements are used to determine liability, recognize rights to use data and to identify restrictions on data usage. U.S. exchanges have tended historically to specify their own Subscriber Agreements. Many overseas exchanges require vendors to ensure that the vendor-subscriber agreement protects the essential interests of the exchanges with regard to their market data. The debate is about the need for these agreements as well as the form and content of the agreement. This includes the creation and use of a common form agreement and the use of click-on agreements for both professionals and non-professionals.

Billing and Reporting Requirements: Current mechanisms for billing and reporting are cumbersome, costly and inefficient for everyone in the industry. For user firms this is about invoice reconciliation and managing their market data inventory. For exchanges this is about getting compensated for market data. Vendors are in the middle, frequently acting as the licensee of the exchange as well as the reconciliation agent for the user. Billing and reporting is multiple-system and very manually intensive. There is broad global agreement on the objectives of simplifying and automating the billing process, but it is a huge task complicated by inconsistencies in reporting requirements and billing approaches (direct versus indirect).

Non-Real Time Data: This refers to the market data policies governing delayed data. There is general concern to ensure that delayed data is not passed off as real-time. However, the contractual status of delayed data varies substantially between exchanges and the practical problems of both determining intellectual property rights and applying rules and policies to data are considerable.

Derived Data: Most market data agreements include general or specific rights for vendors and subscribers to process data. Very few contracts are entirely clear on the extent to which vendors and subscribers are licensed to create or derive their own intellectual property from exchange market data (i.e. indices, graphs or historical analyses) and on the extent to which data derived in this way is subject to market data fees and reporting requirements.

The administrative burdens related to the management and use of market data are substantial and costly to all participants in the market data industry. As a general rule, FISD members believe that the provision of market data should be easy to administer and flexible enough to accommodate new technologies and electronic commerce initiatives.

The good news is that all segments of the industry - exchanges, vendors and user firms alike - agree with this objective and are working together to simplify, standardize and automate business processes. However, there are at least two significant obstacles that must be addressed to make this objective a reality. The first is agreement on the adjustments to business requirements. The second is simplification and automation of business processes to promote efficiency. Let us offer a few examples to illustrate the issues:

In terms of business requirements, the issues associated with direct versus indirect billing and unit of count are good illustrations of the complexities of the factors at work and areas where there is an honest difference of opinion among industry participants.

Direct vs. Indirect Billing

NYSE, OPRA, and Nasdaq (for data feeds) have a direct contractual and billing relationship with their subscribers. They have a prior-approval requirement through the Exhibit A/Schedule A process. Some vendors have argued that this is unnecessary and administratively cumbersome. They would like to see the exchanges adopt market data contracts that authorize the use of data according to the rules specified by the exchange. In fact, this is the business model in use by most of the exchanges around the world.

The North American equity exchanges, on the other hand, have adopted this convention because they say it is needed to support their direct contractual business model as well as their billing requirements. The debate centers on whether the exchanges are willing to migrate to vendor billing. The benefits and drawbacks of vendor billing is a subject of great debate within the industry. The exchanges, and many of the large user firms that report direct to the exchange, do not advocate vendor billing because of concerns about the accuracy of invoices and because of the difficulties of reconciliation associated with indirect billing and reporting processes.

Unit of Count

Standardization of the unit of count is another area of great debate. Do you count units, passwords, or locations? If two data feeds are supporting one user, is that one unit or two? What happens if there are two terminals on one desk -- one for analysis and the other for trading - is that one unit or two? What about for shared terminals, is the unit of count the terminal, unique ID, or the total number of users? For an enterprise license, do you base the unit of count on accesses, terminals, the number of registered representatives, trading turnover or some other metric?

Obtaining consensus on business requirements is not a simple task. FISD facilitates improving the overall business climate by providing a forum for open dialogue on the implication of these business issues to occur. We believe that the work our members are doing in this arena indicates a genuine attempt by all sides of the industry to converge on business requirements that can be applied and enforced efficiently and practically on a worldwide basis. Ideally, these business issues would be managed as legitimate business concerns based on the business realities of the involved parties. However, some of our members feel that the Commission might have to exercise oversight on business policy in areas where the policy requirements could be viewed as an "impediment to commerce" or where resolution of these issues proves intractable.

Business Process Automation

The simplification and automation of business processes to promote electronic commerce and operational efficiency is an area where there are significant opportunities for reductions in the cost of market data management and administration. This is also an area where progress is occurring. FISD's role in business process automation is to facilitate and build consensus via industry working groups and to identify, analyze and promote advances introduced by industry leaders. Industry reference tools such as the Exchange Contract Guide and the Market Data Policy Database are being used as resources in this area.

Industry cooperation has already resulted in a consolidated Exhibit A/Schedule A. NYSE is presently testing an automated data feed request process that will significantly reduce the turnaround time for authorization. This is a significant development and one that will allow for faster response time to marketplace needs.

User firms within FISD have proposed a standard unit of count definition for industry-wide consideration. NYSE among other exchanges around the world have adopted this definition and implemented policies and automated billing and reporting systems that allow users to pay only once if they access the same data from multiple sources.

FISD members and working groups will continue to explore the possibilities for reducing inefficiency where possible using current business models. These may include, for example, the adoption of new technology in administrative systems, the use of common industry billing and customer identification codes to facilitate billing automation, extending the use of "click-on" agreements and using automated online administration procedures to replace hard copy documents.

Definition of Core Data

So far, this paper has highlighted the complexity of global market data administration and identified issues where there may be little benefit, in terms of industry efficiency or investor protection, from direct involvement by national regulators. The discussion concerning the level of information required for market transparency currently taking place among the members of the Advisory Committee are areas where leadership by the Commission, in consultation with the industry, could have significant national and global benefits.

And while this paper is about market data business practices, we recognize the concepts of information competition, the role of exchanges as value-added information providers, the notion of mandates on the factors of production and the potential for changes to the display rule make these issues particularly relevant.

I'd like to emphasize that while FISD is not taking an organizational position on these core issues, we do recognize that there are important market data business practice implications of the following:

• What is the definition of core data? Points raised by information vendors such as Bloomberg (see attached paper) suggest that the definition of core data should include all market data that cannot be derived by reverse engineering. According to Bloomberg, core data would likely consist of the National Best Bid/Offer (NBBO), last-sale and other information required by information vendors to build the informational displays and analytical tools needed by the investment community.

  For example, if an exchange sells a derived product, such as a depth-of-book indicator, should it have to make available to vendors and investors all the constituent data elements that were not otherwise available and were used to create the derived product - such as the depth-of-book data that was used to create the indicator? The underlying issues is whether the exchange should retain rights of access to unpublished data that can be used to create a unique competitive advantage in the market of derived data?

• Should depth-of-book data be mandated? The general perspective within the market data industry is that with the advent of decimalization, market participants need to see complete depth-of-book exposure. To preserve the standards of transparency, some members are suggesting that core data should include access to the depth-of-book (i.e. the highest level of transparency offered by an exchange).
  
  
• Should exchanges compete in the provision of value-added information? To reinforce the March 1 Advisory Committee discussions, we have found no objections to the right of exchanges to compete in the value-added information business. The primary issue of concern raised by the vendors relates to the method of ensuring that there is a level playing field on access to core data. In other words, they are suggesting that the discussion examine methods of ensuring that there is no cross-subsidization so that exchanges bear the same costs as independent entities for core data used to create value-added functionality and analytics. Some members have suggested that the use of Exhibit and Attachment A's and the requirement for pre-authorization puts the exchange in a privileged position of knowing what applications are being proposed. They are concerned that this gives the exchange an advantage to the extent they become competitors of information vendors and users.