The types of data that would have be instantaneously tracked would include, for example, compensation for all custodial or pension management services by an asset management affiliate or arm of the broker-dealer, payroll administration services by a data services affiliate of the broker-dealer, corporate credit cards sponsored or serviced by an affiliate, a checking account or safety deposit box maintained at a banking affiliate, or possibly even compensation such as rebates received for office equipment purchases.⁹ This information would have to be tracked for every corporate payee, not just the ones that are currently the subject of the broker-dealer's research coverage, since there would be no way of knowing what the subject companies will be 12 months hence.

Once a system is put in place, 100 per cent compliance will never be possible. It will only take one instance of a delayed report, an overlooked minor payment in a far-flung affiliate, or a report that is mis-entered due to a division of the issuer using a different name, to undermine the system. Compliance will be especially impractical with regard to more recent minor payments received. A reasonably fool-proof system to identify all payments received from a specific corporate entity would be practically impossible no matter how much time, money and effort was devoted to it, but the fallibility of such a system would be even greater with regard to "real-time" payments or payments received in a current fiscal period.¹⁰

These rules would be especially onerous for broker-dealers that have affiliates abroad. These firms will have to capture on a real-time basis every penny, euro, peso, ruble, dinar, yuan and yen of compensation received around the globe by any of its business lines. This imposes an obligation on any U.S. broker-dealer that conducts research and its affiliates worldwide that is unparalleled in any other jurisdiction.

The costs of a requirement to track every scrap of compensation received by a broker-dealer and its affiliates will be enormous for all firms, both those with affiliates worldwide and those whose affiliates are located only within the United States. The exact dollar amount of cost is difficult to estimate in the brief time during which comments on Amendment No. 2 are being accepted, but it could well be a securities industry-wide cost into the hundreds of millions of dollars to develop such systems, with very substantial costs to operate the systems going forward. This will create a huge impediment to the ability of U.S. broker-dealers and their affiliates to remain competitive with other financial institutions domestically and around the globe that do not face these costs.¹¹

Instead of disclosure that is focused on compensation that poses a true conflict of interest, and that was known or reasonably knowable to the broker-dealer or the analyst, the proposed rules require disclosure of whether the broker-dealer or one of its affiliates received any compensation from the subject company, however minor and regardless of whether the broker-dealer or the analyst involved would otherwise have had any inkling about it. In other words, if an affiliate of the broker-dealer received a dollar of compensation (or foreign currency equivalent) for some very minor service such as renting a safety deposit box or maintaining a checking account, that payment would obligate the broker-dealer to state in its research reports on the subject company words to the effect that "the broker-dealer or its affiliates have received compensation other than for investment banking services from the subject company in the past 12 months."

Disclosure this broad tells investors little to help them appraise the objectivity of broker-dealer research. Investors will have no way of judging from this disclosure if the compensation poses a material conflict of interest. The entire point of the Sarbanes-Oxley requirement, to require disclosure of receipt of any compensation which poses a conflict of interest that is "known or should have been known . . . to exist at the time of the appearance or the date of distribution" is lost under this approach. Moreover, an uneven playing field will have been created. Investors who are well-versed in securities regulation, such as sophisticated institutional investors, will understand that the disclosure is essentially meaningless. Less sophisticated investors, such as most retail customers, may not understand what the disclosure represents, and may assume that every time this disclosure appears it represents a significant conflict of which the broker-dealer and the research analyst had advance knowledge.

Alternative Approaches. We have developed two possible formulations of a rule that we think better embodies what Sarbanes-Oxley directs than do the rules contained in Amendment No. 2.

Our preferred alternative is as follows:

(1) A member must disclose in research reports, and an [associated person] [research analyst] must disclose in public appearances that a conflict of interest may exist due to compensation received by the member or its affiliates if

(a) the [associated person] [research analyst] primarily responsible for the report or making the public appearance knows at the time of the report or appearance of any compensation received by the member or any affiliate of the member within the past 12 months;

(b) the [associated person] [research analyst] primarily responsible for the report or making the public appearance should have known of any compensation received by the member or any affiliate of the member within the past 12 months;

(c) any supervisory analyst or member of a committee with direct influence or control over preparation of research reports knew or should have known about compensation received by the member or any affiliate of the member within the past 12 months.

(2) Paragraphs 1(b) and 1(c) shall be satisfied if the member maintains a system or process requiring all [associated persons] [research analysts], supervisory analysts, and member of committees with direct influence or control over preparation of research reports to disclose on a monthly basis to a designee in the research department all instances in which they are aware of compensation that the member or any of its affiliates has received from a subject company.

(3) A member will not have violated paragraph 1 with regard to any compensation that does not present a conflict of interest.

This rule focuses on information known to the research department. It assumes that the research department is synonymous with the broker-dealer for purposes of the rule. This is logical and consistent with the purpose of this provision, to disclose "conflicts of interest that can also influence the objectivity of the analyst in preparing the report. Senate Report at 33. It also is congruent with the requirement in Sarbanes-Oxley requiring rules that restrict "the prepublication clearance or approval of research reports by . . . persons not directly responsible for investment research. . . ." Exchange Act Sec. 15D(a)(1)(A). This proposal also uses, and is consistent with, Sarbanes-Oxley's express requirement of disclosure of conflicts of interest that are "known or should have been known." A rule triggering disclosure based on compensation that is known to one or more persons who have no responsibility for investment research would not be "reasonably designed" to require disclosure of "conflicts of interest," since those persons are barred by other rules from having input into the content of the research report, and therefore any compensation that they know or should know about does not constitute a conflict of interest with regard to research reports.¹³

The first trigger for this rule would be whether those responsible for a research report actually know about any compensation received by the broker-dealer or its affiliates at the time the report is issued. This is the core potential conflict that Exchange Act Section 15(b)(2) is intended to address.

Second, the rule would capture any other compensation constituting a conflict of interest that is known to others in the research department or that the research department or the analyst should know about. This would not attempt to capture all information about compensation conflicts on a real-time basis, since that would involve canvassing many people (many thousands in the case of some firms) every time a research report is issued or an analyst makes a public appearance, resulting in delays in issuing the report that could disadvantage investors. Sarbanes-Oxley's requirement for a "reasonable design," as well as the Commission's exemptive authority under the Exchange Act, allows for writing the rule to take this impracticality into account.

A workable approach would be to require firms to establish systems to require analysts and research management to disclose to a designated person within the research department on a periodic basis whether they are aware of any compensation received by the member or any of its affiliates. This is reflected in paragraph 2 of our proposal. Under our proposal an analyst or a member of research management would have an obligation to make a monthly report to this designated individual of all compensation that they know or "should have known" about. The "should have known" standard is intended to reflect the scienter standard of actual knowledge or reckless disregard that is well-developed in case law under SEC Rule 10b-5. See, e.g., Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193 (1976). One point is of critical importance. This provision should not be interpreted to require anyone responsible for research to search out information about compensation that they would not otherwise obtain in the ordinary course of their work.¹⁴ As noted at pages 10-11 above, the point of the rule is lost if it requires analysts to inform themselves about potential conflicts that they would otherwise not know about. Express guidance from the regulators clarifying this point, in the adopting release or in interpretive guidance, would be helpful.

Finally, our proposal includes a clause reflecting the legislative language that only compensation constituting a conflict of interest should be included in the rule. The statute directs that the rule should be designed to require disclosure of "conflicts of interest" arising from receipt of compensation that are known or knowable to the broker-dealer or the securities analyst. Therefore, if a given item did not constitute a conflict of interest it should not violate the rule.¹⁵ It is appropriate to include a provision setting out such a defense.

Our other suggested formulation is:

(1) A member must disclose in research reports, and an [associated person] [research analyst] must disclose in public appearances that a conflict of interest may exist due to compensation received by the member or its affiliates if

(a) the [associated person] [research analyst] primarily responsible for the report or making the public appearance knows at the time of the report or appearance of any compensation received by the member or any affiliate of the member within the past 12 months; or

(b) any compensation that has been identified as received from the subject company by the member or any of its affiliates in accordance with paragraph (2).

(2) The member shall have made a sufficient determination of whether any compensation was received under (1)(b) if the member has taken steps reasonably designed to identify within 60 days after completion of the member's last fiscal year whether any compensation was received from the subject company in the ordinary course of business in that fiscal year, provided that paragraph (1)(b) shall not apply to any subject company as to which the member initiated coverage since the beginning of the member's current fiscal year. In any such instance, the research report must state or the [associated person][research analyst] must disclose in public appearances, that the member or its affiliates may have received compensation from the subject company in addition to any other compensation required to be disclosed.

(3) For purposes of this section, the term "affiliates" has the same meaning as "covered persons" under Regulation AC, 17 C.F.R. Sec. 242.500.

(4) A member will not have violated paragraph 1 with regard to any compensation that does not present a conflict of interest.

(1)(a) of this proposal is identical to (1)(a) of our first proposal, and paragraph (4) of this proposal is identical to paragraph (3) of our first proposal. There is a significant difference in (1)(b), which would require the broker-dealer to track compensation received by the member and its affiliates. This requirement is modified by paragraphs (2) and (3) to take into account the many problems identified on pages 6-10 above that a requirement to track all compensation on a real-time basis poses for firms and for investors. Paragraph (2) establishes that the member must make a due diligence effort within 60 days after the end of every fiscal year of the member to identify every covered company as to which the firm or one of its affiliates received compensation in the broker-dealer's last fiscal year, and the fact that such compensation was received must be disclosed in research reports and public appearances about the subject company for the following fiscal year if the compensation constitutes a conflict of interest.

Companies as to which research is newly initiated pose a special problem. The member would have had no reason to search for compensation from those companies when conducting its annual review. In these limited circumstances, we propose in paragraph (2) a standardized disclosure that research recipients or public appearance attendees should assume that the broker-dealer or its affiliates may have received compensation from the subject company in addition to any that is otherwise disclosed. We note that this is a much narrower use of standardized disclosure than occurred in the Global Settlement, where the SROs and the Commission recognized the prophylactic value of standardized disclosure, requiring that a similar but much broader standardized disclosure should be included in all research reports issued by the settling firms.¹⁶ In light of the enormous practical problems that ensue from trying to track all compensation received from a company that there has not heretofore been a reason to track, we think this limited use of standardized disclosure is an appropriate step "reasonably designed" to meet the Sarbanes-Oxley requirements.

Paragraph (3) would link the scope of the affiliates swept up into the search for compensation to the concept in SEC Regulation AC of "covered persons." Just as in Regulation AC, it makes sense to restrict the search for compensation to affiliates that have "officers or employees in common with the broker or dealer who can influence the activities of research analysts or the content of research reports," or that lack "policies and procedures reasonably designed to prevent the broker or dealer, any controlling persons, officers . . . and employees of the broker or dealer from influencing the activities of research analysts and the content of research reports . . . ." 17 C.F.R. Sec. 242.500. Requiring tracking and disclosure of compensation received by affiliates that have no influence over the content of research reports or analysts' public appearances would add significant expense, while leading to potentially misleading disclosures.

a. Definition of "Client." NYSE Rule 472(k)(1)(i)(c) and NASD Rule 2711(h)(2)(F) require a broker-dealer to disclose in research reports, and a research analyst to disclose in public appearances (if the analyst knows or has reason to know), that the subject company is or has been "a client" of the broker-dealer in the preceding 12 months. This is intended to fulfill the statutory requirement to adopt a rule "reasonably designed" to disclose whether a subject company "currently is, or during the 1-year period preceding the date of the report has been, a client of the registered broker or dealer, and if so stating the types of services provided to the issuer." Exchange Act Sec. 15D(b).

The term "client" needs to be precisely defined in order for firms to design a consistent and objective way of producing the required information.¹⁷ However client is defined, firms need guidance on how to go about evaluating whether a subject company was a "client" 12 months ago. How should a firm classify a subject company if the firm had some sort of business relationship with the company more than 12 months ago, but that relationship has been neither active nor formally terminated within the past 12 months? Firms also need help in understanding how to determine whether a subject company is a "client" as of the date of the research report or public appearance, in light of the practical problems that we describe above about tracking fluid and highly dispersed information on a real-time basis (albeit without having to worry about affiliates in this context). In addition, the rule should be crafted so that it does not require research analysts to search out for information about potential conflicts that they would not otherwise know about.

The answers to these questions are bound to be somewhat arbitrary. What is critical is that the regulators select a consistent standard, so that firms are not left to rely solely on their own subjective, and differing, approaches to compliance with this requirement. We suggest that, since the Act treats compensation and client disclosure similarly, the term "client" be defined in terms of compensation received by the broker-dealer. Specifically, a firm could identify compensation received by the broker-dealer from the subject company in the same manner as we suggest for tracking compensation received by the broker-dealer and its affiliates under our second proposal regarding tracking compensation (pages 15-16 above). The rule might be phrased as follows:

(1) A member is required to disclose in a research report, and [an associated person] [a research analyst] is required to disclose in a public appearance, that the subject company is a client of the member, if

(a) the [associated person] [research analyst] primarily responsible for the report or making the public appearance knows at the time of the report or appearance of any compensation received by the member or any affiliate of the member within the past 12 months; or

(b) any compensation has been identified as received by the member or any of its affiliates in accordance with paragraph (2).

(2) The member shall have made a sufficient determination of whether any compensation was received under (1)(b) if the member has taken steps reasonably designed to identify within 60 days after completion of the member's last fiscal year whether any compensation was received by the member from the subject company in the ordinary course of business in that fiscal year, provided that paragraph (1)(b) shall not apply to any subject company as to which the member initiated coverage since the beginning of the member's current fiscal year. In any such instance, the research report must state or the [associated person][research analyst] must disclose in public appearances, that the subject company may have been a client of the member.

(3) Any disclosure required under paragraph (1) above shall also disclose the types of services provided to the subject company. For purposes of this paragraph, the types of services provided to the subject company may be described as investment banking services, non-investment banking securities-related services, and non-securities services, provided that disclosures required under paragraph 1(a) need include information as to the type of services provided only to the extent actually known to the [associated person][research analyst].

(4) A member will not have violated paragraph 1 with regard to any compensation that does not present a conflict of interest.

Coupled with the current mandated disclosure of any compensation received from the subject company for investment banking services, ¹⁸ and the other new disclosures that we propose above regarding compensation, this would provide investors with an extremely useful and comprehensive view of potential conflicts of interest that might be significant to their weighing of the research report or public appearance, while addressing the compliance concerns that arise absent any guidance on how to determine if a subject company has been a "client" in the past 12 months. We recognize that our proposal is not perfect. It may identify some "client" relationships, due to compensation received well over a year ago, which would not otherwise be thought to give rise to a client relationship in the past 12 months. It may also miss some client relationships that commenced in the past 12 months, if the analyst writing the research report or making a public appearance is unaware of them. The statute, however, does not demand literal-mindedness, but rather a rule "reasonably designed" to fulfill its objectives. We think that our proposal does this far more effectively than the requirement set out in Amendment No. 2.

c. Definition of "Subject Company." The SROs have not made clear whether "subject company" includes related entities such as subsidiaries, parent companies, and pension or other, similar funds that may be affiliated. Including subject company affiliates will make tracking systems even more complex and unworkable. Sarbanes-Oxley does not require inclusion of related entities. If that term does extend to related entities, then in some cases a firm or analyst will not know that a related entity is related, and will fail to make required disclosures. In other cases a firm or analyst will infer that an entity may be related to an issuer and, for safety's sake, make disclosures when no potential conflict of interest actually exists. We recommend that the SROs clarify that "subject company" is limited in accordance with NYSE Rule 472.60 and NASD Rule 2711(a)(9) (i.e., the actual "company whose equity securities are the subject of research reports"). With such clarification, this rule would be more consistent with the express language of Sarbanes-Oxley, which focuses on compensation from, and a client relationship with, the "issuer that is the subject of the appearance or research report" (see Sarbanes-Oxley § 501(a), Exchange Act §15D(b)(2) and (3)).

472 (g)(2) and 2711(j): antiretaliation provision.

This provision is compelled by Sarbanes-Oxley and the language closely tracks the statute. See Exchange Act Sec. 15D(a)(1)(C). We strongly agree with the authors of the Act that analysts should not face retaliation for writing a negative research report, so long as the report reflects the analysts' own view. We are concerned that in some instances an underperforming analyst might try to use this provision as a ploy to protect himself or herself from consequences for misconduct or poor performance by putting out one or more spurious negative reports. While the rule language and the statutory provision on which it is based seek to address this point, it would be helpful if the SEC and both SROs underscored in the discussion in the adopting release that an analyst does not immunize himself or herself from adverse employment action simply by putting out a negative research report.

472(f)(2) and (3), 2711(f)(1) and (2): bar on public appearances by lead and co-managers, and 25-day bar on syndicate and participant and dealer research.

Extension of the quiet periods to dealers who participate in an underwriting is required by Sarbanes-Oxley, but other aspects of Amendment No. 2's proposed changes to the quiet periods are not. See Exchange Act Sec. 15D(a)(2). Nevertheless, we have no objections to the proposed changes. We request clarification in one respect: whether the exception in 472(f)(2) and 2711(f)(1) for research reports permitted under SEC Rule 139 also applies to public appearances concerning issuers for which Rule 139 research is permitted.

B. Other Changes.

Rule 472 Interpretation on Communications with the Public.

This clarification is helpful and we support it. The NASD made a similar statement in its proposing release for Amendment No. 1, and the two SROs now seem to be in basic agreement on this point.

In response to the request for comment on whether the record-keeping requirement in the Interpretation is appropriate, we think that the NYSE's proposed record-keeping requirement is unnecessary, since firms will find that they need to keep these records in any event to demonstrate compliance. Moreover, we have substantial concerns with the record-keeping requirement as proposed by the Exchange. This provision differs substantially from the record-keeping requirements imposed by the NASD. We do not understand why this must be framed as a nondelegable obligation on the securities analyst. We also question why the Exchange insists that such a record must be created prior to opening for business on the next business day. These requirements will be extremely difficult for a research analyst to personally meet in many situations, such as where an analyst is traveling, or speaks to a reporter just prior to the market opening, and therefore is unable to personally create the record in the allotted time.

It should be sufficient to require that the research analyst who conducted the interview or made the appearance, or a legal or compliance official assigned to the research department, prepared such a record or caused such a record to be prepared at the earliest practicable time by an appropriate officer or employee of the firm. If a requirement like this is included in the rules as adopted, we also ask that the NYSE and NASD clarify that, in light of the record-keeping requirement, the obligation to "track" whether disclosures were in fact made by the media outlet is unnecessary.

NYSE 472(b)(5) "No member or member organization may publish ..."

The NYSE proposal makes an interesting change in its proposed "pitch rule" regarding restrictions on the ability of research analysts to write if they are have communications "in furtherance of obtaining investment banking business." As we discussed in our March 10 and May 9 letters, we think that this provision is seriously flawed, unworkably vague, and inconsistent with the requirements of the Term Sheet.¹⁹ The NYSE's proposed change, which it does not explain, would change the rule to bar publication of research by "the member or member organization" rather than by the research analyst, if the analyst communicates with the subject company "in furtherance of obtaining investment banking business." We are not sure of the reason for this change, which the NASD proposal does not make, but we note that it marks another instance of the two rules diverging from each other, and we do not believe that it meaningfully addresses the concerns that we have raised about this proposal.

472(m) and 2711(k): Small firm exception.

We support this limited carve-out for small firms from some of the structural requirements of the rules. The parameters of the definition of "small firm" are very narrow, and few if any firms that actually conduct research and investment banking activities will fit in them. It would be helpful if the SROs could clarify that the term "investment banking services transactions" for this purpose does not include municipal securities transactions.

C. Implementation Periods.

As noted above, the SROs' proposed requirement for firms and their affiliates to build and test worldwide systems to track on a real-time basis all compensation received by every company that might potentially become a subject of the firm's research in the next year will be a tremendous challenge. It will be essential to have substantially more than 120 days to create these systems. If the SROs' proposal is adopted in its current form, firms will need very substantial time to examine whether their current internal systems can be enhanced to try to meet the rules' requirements, and which systems will have to be replaced entirely. Firms will then have to assess the continuing viability of offering research in light of those costs. Once these determinations are made, firms that decide to continuing offering research will need much more time to make the necessary enhancements, or to design and build and test entirely new systems. We propose 18 months from the adoption date as a more realistic timeframe than the proposed 120 days. If our first suggestion, on pages 11-12 above, is adopted, we believe that the current proposed implementation period is appropriate. If our second suggestion, on pages 15-16 above, is adopted, significant time will be required for firms to build systems and/or procedures to identify compensation received by the broker-dealer or its affiliates in the last fiscal year. Therefore, we suggest an implementation period of 12 months from the date of adoption. For our recommendation regarding disclosure of client relationships with subject companies we also recommend a 12 month implementation period.

Extension of the trading restrictions to research management, if adopted, should also permit more than a 60 day implementation period to permit affected persons to wind down positions, restructure portfolios, etc. and for firms to redesignate personnel to oversee research in the likely event that some individuals will seek to remove themselves from research supervisory committees.²⁰ We propose 180 days for implementation of these requirements.

D. Additional Suggestions in Light of Global Settlement.

In response to the request additional comment on the SRO rule changes that were proposed in the Original Notice in light of the Global Settlement, we have one additional point to add to the ones we made in our May 9 letter. There is a notable disparity between the proposed amendments to the SRO rules and the Global Settlement regarding withdrawal of research coverage of a subject company. As described in the Original Notice, the SROs would require that when firms "withdraw" research coverage of a subject company they would be required to provide notice of the withdrawal in the same manner as when research coverage was first initiated by the firm and must include the firm's final recommendation or rating. The Global Settlement has a similar requirement, but excludes reports from the provision "for any company as to which the firm's prior coverage has been limited to purely quantitative analysis."²¹ It would seem appropriate for the SRO provisions to contain a similar exclusion

Conclusion.

We hope that these comments on Amendment No. 2 to the proposed SRO rules, together with the more substantial suggestions of our March 10 and May 9 letters, will result in a more workable set of rules that will help to improve public trust and confidence in research analysts. Because of the complex implementation challenges that are raised by the proposal, if the SROs and the SEC wish to proceed with anything like the current proposal, prior to setting an implementation schedule we urge the staffs to meet with operational and business officials from broker-dealers so that they can have a better understanding of the implementation challenges that the rules will pose. We would be happy to organize such a meeting.

If you have any questions on any aspect of this letter, please contact George R. Kramer, staff adviser to the Committee, at 202-296-9410, or by e-mail to gkramer@sia.com.

Sincerely,

Robert C. Dinerstein, Chairman
SIA Federal Regulation Committee

Attachments:

March 10, 2003 comment letter

May 9, 2003 comment letter

Cc (w. attachments):

Chairman William H. Donaldson, U.S. Securities and Exchange Commission

Commissioner Paul S. Atkins, U.S. Securities and Exchange Commission

Commissioner Roel C. Campos, U.S. Securities and Exchange Commission

Commissioner Cynthia A. Glassman, U.S. Securities and Exchange Commission

Commissioner Harvey J. Goldschmid, U.S. Securities and Exchange Commission

Robert R. Glauber, Chairman and Chief Executive Officer, National Association of Securities Dealers, Inc.

Mary L. Schapiro, President, NASD Regulation, Inc.

Elisse B. Walter, Chief Operating Officer and Executive Vice President, NASD Regulation, Inc.

Thomas Selman, Senior Vice President, NASD Regulation, Inc.

Philip A. Shaikun, Assistant General Counsel, NASD Regulation , Inc.

Richard Grasso, Chairman and Chief Executive Officer, New York Stock Exchange, Inc.

Edward A. Kwalwasser, Group Executive Vice President, New York Stock Exchange, Inc.

Donald van Weezel, Vice President, Regulatory Affairs, New York Stock Exchange, Inc.

Annette Nazareth, Director, Division of Market Regulation, U.S. Securities and Exchange Commission

Alan Beller, Director, Division of Corporation Finance, U.S. Securities and Exchange Commission

Stephen Cutler, Director, Division of Enforcement, U.S. Securities and Exchange Commission

Giovanni P. Prezioso, General Counsel, U.S. Securities and Exchange Commission

Robert L. D. Colby, Deputy Director, Division of Market Regulation, U.S. Securities and Exchange Commission

Larry L. Bergmann, Senior Associate Director, Division of Market Regulation, U.S. Securities and Exchange Commission

James A. Brigagliano, Assistant Director, Trading Practices, Division of Market Regulation, U.S. Securities and Exchange Commission

Thomas D. Eidt, Division of Market Regulation, U.S. Securities and Exchange Commission

Racquel L. Russell, Division of Market Regulation, U.S. Securities and Exchange Commission

____________________________

1	The Securities Industry Association brings together the shared interests of more than 600 securities firms to accomplish common goals. SIA member-firms (including investment banks, broker-dealers, and mutual fund companies) are active in all U.S. and foreign markets and in all phases of corporate and public finance. Collectively they employ more than 495,000 individuals, representing 97 percent of total employment in securities brokers and dealers. The U.S. securities industry manages the accounts of nearly 93-million investors directly and indirectly through corporate, thrift, and pension plans. In 2001, the industry generated $280 billion in U.S. revenue and $383 billion in global revenues. (More information about SIA is available on its home page: http://www.sia.com.) SIA's Federal Regulation Committee, consisting of over two-dozen chief legal officers of securities firms, is SIA's principal legal and regulatory committee.
2	We commented extensively on March 10, 2003 on the initial version of the proposed revisions to the SROs' rules relating to research conflicts (copy attached). In that letter we expressed support for a number of the proposed changes, but also pointed to a number of inconsistencies between the two SROs' sets of rules, ambiguities in the rules, and likely unintended consequences of some of the proposals that could weaken rather than strengthen investor protections, or make the rules unworkable in many respects. We supplemented our comments on May 9, 2003, in response to a settlement between the Securities and Exchange Commission ("SEC"), the SROs, and a number of state authorities and ten major broker-dealers following an investigation into research analyst conflicts of interest. See http://www.sec.gov/litigation/litreleases/finaljudgadda.pdf (the "Global Settlement")(copy attached). Our letter called attention to a number of respects in which the obligations placed on the settling firms in the enforcement action are easier to understand or to apply than the requirements of the current or proposed SRO rules concerning research conflicts.
3	Regarding the invitation in Amendment No. 2 to submit comments on the original proposed SRO rule changes in light of the Global Settlement, we supplement our May 9 letter with an additional comment at page 26 below.
4	We assume that the term "affiliate" will continue to be defined by the SROs to include "any company that controls, is controlled by, or is under common control with, the member" and will not include individuals employed by the member or one of its affiliates. See Joint Memorandum of NASD and the New York Stock Exchange, NASD NTM02-39, July 2002, ("Joint Interpretive Memorandum") at 374.
5	This plain reading of the words of the statute is backed up by the legislative history of the Act. This provision of the Act originated in the Senate Committee on Banking, Housing and Urban Affairs. The Committee Report stated that this provision "is intended to . . . provide disclosure to investors of certain conflicts of interest that can also influence the objectivity of the analyst in preparing a research report." Report of the Committee on Banking, Housing and Urban Affairs to Accompany S. 2673, Rpt. 107-205, June 26, 2002, at 33 (emphasis added) ("Senate Report").
6	In addition, other provisions of the Exchange Act direct that the SRO rulemaking process take into account the public interest, as well as promoting competition. Exchange Act 3(f) requires the SEC to consider, as part of its weighing of the public interest in reviewing a proposed rule of a self-regulatory organization, "whether the action will promote efficiency, competition and capital formation." The NASD and NYSE's rulemaking authorities are under similar obligations. Section 15A(6) of the Exchange Act requires that any rule of the NASD "remove impediments to and perfect the mechanism of a free and open market . . . and, in general, to protect investors and the public interest." Exchange Act Section 6(b)(8) requires that the NYSE's rules "not impose any burden on competition not necessary or appropriate in furtherance of the purposes of this title."
7	Essentially all of the major broker-dealers in the United States have non-broker-dealer affiliates.
8	Based on NYSE Focus data and data about SIA member firms, well in excess of 80 per cent of the broker-dealer business in the U.S. is conducted by firms that have foreign affiliates.
9	As discussed at footnote 15 below, we hope that the regulators will not view as "compensation" types of payment which on their face pose no danger of conflicts of interest, such as ordinary-course-of-business rebates.
10	If the subject company has a customer account with the broker-dealer or an affiliate, anti-money-laundering rules might require real-time tracking of any funds coming into the account. However, account deposits are surely not what Congress meant by "compensation." In any event, this would hardly capture every type of compensation that the firm or its affiliates might receive from the subject company.
11	As noted at footnote 6 above, the SRO rulemaking process is supposed to have as one of its aims promoting competition.
12	This requirement will be especially awkward under the NYSE's proposed rule amendments, because proposed amendments to Rule 472(l) requires that these disclosures be made by any employee who discusses specific securities or industries in a public appearance, regardless of whether that employee is in any way an "associated person" who is involved in preparing research reports. Page 28 of our March 10, 2003 letter discusses our concerns with this provision.
13	This is also consistent with the view expressed in the existing SRO research rules that members should disclose in research reports or public appearances "any other actual, material conflict of interest . . .of which the research analyst knows or has reason to know. . . ." NYSE Rule 472(c) and NASD Rule 2711(h)(1)(C) (emphasis added).
14	This is true whether or not the information is arguably in the public domain. There should not be a continual obligation on the part of analysts or research management to conduct repeated web searches or other searches of public domains to try to locate information about compensation that they would not otherwise become aware of in the course of their research.
15	Examples of a source of `compensation" that should not pose a conflict of interest would be performance on a fully guaranteed contractual commitment to pay, such as a repayment by a subject company of principal on a fully collateralized loan, or a rebate received on an office product purchase or other ordinary-course purchase. We hope that payments such as these would not be deemed "compensation" by the regulators, but even if they are, it is hard to see under any analysis how they give rise to a conflict of interest. There are likely to be other situations where, under the particular facts, no reasonable person could view a form of compensation as posing a conflict of interest.
16	The Global Settlement term sheet required a standard disclaimer that "[Firm] does and seeks to do business with companies covered in its research reports. As a result, investors should be award that the firm may have a conflict of interest that could affect the objectivity of this report." See Addendum A to Settlement, at 9, http://www.sec.gov/litigation/litreleases/finaljudgadda.pdf.
17	Moreover, there may be situations, such as some institutional brokerage subaccounts, where a broker-dealer may not know the identity of all potential "clients." In the institutional brokerage business, "clients" tend to be professional money managers, who generally open sub-accounts for their clients and, in some cases, do not reveal the names of the subaccount owners. These subaccounts pay commissions to the member firm based on trades directed by the professional money manager. If the SROs consider subaccount owners to be "clients" of a broker-dealer, then a firm would be obligated to look for subaccounts in the name of the subject company, but could not do that if the firm does not know the identity of all subaccount owners.
18	Joint Interpretive Memorandum, NTM 02-39, at 379.
19	See pages 11-18 of our March 10, 2003 letter and pages 5-6 of our May 9, 2003 letter (both attached).
20	Moreover, if trading restrictions are imposed on research review committees, there may be uncertainty about whether members of research review committees who have responsibilities for discretionary trading, such as private client asset management, or the firm's proprietary trading, will face trading restrictions on that trading (i.e., is control over proprietary or discretionary trading for the firm or its clients covered by the restrictions on trading by "associated person or member of the associated person's household"?). Firms that take a cautious interpretation will need time to find ways of managing this type of trading in light of this restriction. The SROs could eliminate this concern by giving an interpretation that the trading restrictions only apply to an associated person's personal trading.
21	Global Settlement, Addendum at 6, http://www.sec.gov/litigation/litreleases/finaljudgadda.pdf.