Securities Arbitration: Actions Needed to Address Problem of Unpaid Awards (Chapter Report, 06/15/2000, GAO/GGD-00-115). Pursuant to a congressional request, GAO provided information on issues relating to the arbitration process in the securities industry, focusing on: (1) whether arbitration forums had implemented recommendations made in GAO's 1992 report and whether the changes were effective; (2) how investors fared in award decisions; and (3) the extent to which investors were paid the amounts awarded by arbitration panels. GAO noted that: (1) the securities industry self-regulatory organizations (SRO)--the National Association of Securities Dealers (NASD) and the New York Stock Exchange (NYSE)--implemented GAO's 1992 recommendations by giving arbitration participants a larger role in selecting arbitrators, periodically surveying arbitrators to verify background information, and improving arbitrator training; (2) NASD and NYSE generally assessed arbitrator performance using methods such as participant evaluations, focus groups, and discussions with arbitration attorneys; (3) NASD and NYSE believes that this participant feedback indicated that the program changes have improved investors' perceptions of arbitrator performance and the fairness of the arbitration process; (4) GAO could not reach conclusions about the fairness of the arbitration process from case outcome statistics; (5) investors did not receive as high a percentage of favorable arbitration awards during any year from 1992 through 1998, 49 to 57 percent, as they had during the period of January 1989 through 1990, 59 percent; (6) the percentage of the amount claimed that was awarded also declined during this period; (7) however, an increase in the percentage of cases settled during this period, generally 50 to 60 percent of the total cases concluded, may have changed the mix of cases going to a final arbitration award decision; (8) GAO's survey of investors who received awards in 1998 found that a number of broker-dealers that had left the securities industry often did not pay arbitration awards rendered against them; (9) nearly all unpaid awards were from cases decided in NASD's arbitration forum; (10) when investors complained, NASD took action to suspend nonpaying broker-dealers and had success in recovering awards; (11) most of the unpaid awards resulted from broker-dealers that were no longer in business; (12) NASD is considering changes to its processes to reduce costs and increase options for investors, but these changes still may not address the problem of defunct brokers not paying awards; (13) ultimately, recovering losses caused by undercapitalized, financially irresponsible, or unscrupulous broker-dealers is difficult, if not impossible, for investors; and (14) educating investors about the risks of doing business with such broker-dealers could help them avoid situations that may result in unpaid awards. --------------------------- Indexing Terms ----------------------------- REPORTNUM: GGD-00-115 TITLE: Securities Arbitration: Actions Needed to Address Problem of Unpaid Awards DATE: 06/15/2000 SUBJECT: Securities arbitration Brokerage industry Arbitrators Administrative remedies Stock exchanges Securities regulation ****************************************************************** ** This file contains an ASCII representation of the text of a ** ** GAO report. Delineations within the text indicating chapter ** ** titles, headings, and bullets are preserved. ** ** ** ** No attempt has been made to display graphic images, although ** ** figure captions are reproduced. Tables are included, but ** ** may not resemble those in the printed version. ** ** ** ** Please see the PDF (Portable Document Format) file, when ** ** available, for a complete electronic file of the printed ** ** document's contents. ** ** ** ** A printed copy of this report may be obtained from the GAO ** ** Document Distribution Center. For further details, please ** ** send an e-mail message to: ** ** ** **** ** ** ** with the message 'info' in the body. ** ****************************************************************** GAO/GGD-00-115 United States General Accounting Office GAO Report to Congressional Requesters June 2000 GAO/GGD-00-115 SECURITIES ARBITRATION Actions Needed to Address Problem of Unpaid Awards Ordering Copies of GAO Reports The first copy of each GAO report and testimony is free. Additional copies are $2 each. Orders should be sent to the following address, accompanied by a check or money order made out to the Superintendent of Documents, when necessary. VISA and MasterCard credit cards are accepted, also. Orders for 100 or more copies to be mailed to a single address are discounted 25 percent. Order by mail: U.S. General Accounting Office P.O. Box 37050 Washington, DC 20013 or visit: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW) U.S. General Accounting Office Washington, DC Orders may also be placed by calling (202) 512- 6000 or by using fax number (202) 512-6061, or TDD (202) 512-2537. Each day, GAO issues a list of newly available reports and testimony. To receive facsimile copies of the daily list or any list from the past 30 days, please call (202) 512-6000 using a touch-tone phone. A recorded menu will provide information on how to obtain these lists. Viewing GAO Reports on the Internet For information on how to access GAO reports on the INTERNET, send e-mail message with "info" in the body to: [email protected] or visit GAO's World Wide Web Home Page at: http://www.gao.gov Reporting Fraud, Waste, and Abuse in Federal Programs To contact GAO FraudNET use: Web site: http://www.gao.gov/fraudnet/fraudnet.htm E-Mail: [email protected] Telephone: 1-800-424-5454 (automated answering system) United States General Accounting Office Washington, D.C. 20548-0001 Official Business Penalty for Private Use $300 Address Correction Requested Bulk Rate Postage & Fees Paid GAO Permit No. G100 (233590) Contents Page 101GAO/GGD-00-115 Actions Needed to Address Problem of Unpaid Awards Executive Summary 4 Chapter 1 14 Introduction How Arbitration Is Designed to Work 14 SROs Administer and SEC Oversees 15 Securities Arbitration Objectives, Scope, and Methodology 15 Chapter 2 18 SROs Have Made Changes to Improve Their Arbitration Programs SEC Has Worked With SROs to Improve 18 Securities Arbitration SROs Have Made Changes to Improve 18 Their Arbitration Processes SROs Seek Participants' Views to 21 Assess Effects of Program Changes Chapter 3 23 Arbitration Awards Favoring Investors Declined, but More Cases Were Settled Investors Have Not Fared as Well in 23 Award Decisions as They Did Before 1992, but More Cases Were Settled Factors Affecting Awards for Investors 25 Other Comparisons in Arbitration Case 28 Processing Use of AAA Has Declined 29 Few Securities Disputes Are Litigated, 32 Most Are Dismissed Chapter 4 33 Many Arbitration Awards Were Not Paid Over Half of Awardees Were Not Paid 33 NASD Did Not Routinely Determine If 36 Arbitration Awards Were Paid NASD Is Considering Changes to Address 38 Award Nonpayment by Broker-Dealers That Are No Longer in Business Other Potential Solutions to the 39 Problem of Unpaid Awards Chapter 5 44 Conclusions and Recommendations Conclusions 44 Recommendations 44 Agency Comments and Our Evaluation 45 Appendixes Appendix I: Methodology for Analyzing 50 Arbitration Results Appendix II: Awardee Survey 60 Methodology Appendix III: Awardee Survey 64 Questionnaire Appendix IV: Comments From the NASD 66 Regulation, Inc. Appendix V: Comments From the New York 73 Stock Exchange Appendix VI: Comments From the 75 Securities and Exchange Commission Appendix VII: GAO Contacts and Staff 78 Acknowledgments Tables Table 3.1: Percentage of Cases Decided 26 in Favor of Investors in 1998 Table 3.2: Size and Range of 28 Investors' Claim Amounts in 1998 Table 3.3: Average Processing Time 29 for Cases Decided in 1998 (Days) Table 3.4: Firms Requiring Predispute 30 Arbitration Clauses in Accounts for Individual Investors as of July 1999 Table 3.5: AAA Investor-Initiated, 31 Securities-Related Awards and Results for Investors, 1992 Through 1998 Table 4.1: Estimates of the Percentage 34 of Monetary Customer Arbitration Awards Issued by All Forums and NASD in 1998 That Broker-Dealers or Individual Brokers Paid Nothing, Paid Partially, or Paid Fully Table 4.2: Estimates of Percentage of 34 Award Cases Paid in Full and Dollar Amounts Paid in 1998, by Size of Award Table I.1: Arbitration Cases Won and 54 Lost, Odds on Winning Derived From Them, and Odds Ratios Indicating the Effects of Various Characteristics of Arbitration Cases on the Odds on Winning Table I.2: Arbitration Cases Won That 58 Received Awards That Were More or Less Than 50 Percent of the Amount Claimed, and Odds Ratios Indicating the Effects of Various Claim Characteristics on Winning Sizable Awards Table II.1: Disposition of Sample 61 Figures Figure 3.1: Percentage of Decisions 24 Favoring Investors and Awards to Amounts Claimed, January 1989 Through June 1990 and January 1992 Through December 1998 Abbreviations AAA American Arbitration Association ICS Investors Compensation Scheme NASD National Association of Securities Dealers NASDR NASD Regulation, Inc. NYSE New York Stock Exchange SAC Securities Arbitration Commentator SEC Securities and Exchange Commission SIA Securities Industry Association SICA Securities Industry Conference on Arbitration SIPC Securities Investor Protection Corporation SRO self-regulatory organization B-281372 Page 2GAO/GGD-00-115 Actions Needed to Address Problem of Unpa id Awards B-281372 June 15, 2000 The Honorable John D. Dingell Ranking Minority Member, Committee on Commerce House of Representatives The Honorable Edward J. Markey House of Representatives This report responds to your July 30, 1998, and September 23, 1998, requests that we evaluate issues relating to the arbitration process in the securities industry. On the basis of your requests, this report discusses (1) whether arbitration forums had implemented recommendations made in our 1992 report, Securities Arbitration: How Investors Fare (GAO/GGD-92-74, May 11, 1992) and assessed the effectiveness of the changes; (2) how investors fared in securities arbitration award decisions; and (3) the extent to which investors were paid the amounts awarded by arbitration panels. This report includes recommendations to the Chairman, U.S. Securities and Exchange Commission, regarding regulatory actions to address the problem of unpaid arbitration awards. As agreed with your offices, unless you publicly release its contents earlier, we plan no further distribution of this report until 30 days from its issue date. At that time, we will provide copies to Representative Tom Bliley, Chairman, House Committee on Commerce; Representative Michael G. Oxley, Chairman, and Representative Edolphus Towns, Ranking Minority Member, Subcommittee on Finance and Hazardous Materials, House Committee on Commerce; the Honorable Arthur Levitt, Chairman, U.S. Securities and Exchange Commission; Mr. Frank Zarb, Chairman, National Association of Securities Dealers; Mr. Richard Grasso, Chairman, New York Stock Exchange; and other interested parties. We will also make copies available to others upon request. If you have any questions on matters discussed in this report, please call me or Michael Burnett at (202) 512-8678. Other major contributors to this report are acknowledged in appendix VII. Thomas J. McCool Director, Financial Institutions and Markets Issues Executive Summary Page 7GAO/GGD-00-115 Actions Needed to Address Pro blem of Unpaid Awards Purpose The securities industry uses arbitration to resolve disputes between industry members and individual investors that involve hundreds of millions of dollars each year. Congress, state regulators, and investor groups have questioned whether an arbitration system that is administered largely by securities self-regulatory organizations (SRO) is fair and impartial. In a 1992 report on arbitration, GAO found no indication of a proindustry bias, but concluded that SRO-sponsored forums lacked internal controls to provide investors with reasonable assurance that arbitrators were independent and competent.1 GAO recommended ways for the industry to improve arbitrator selection, qualifications, and training. Because of their continuing concerns about SRO-sponsored arbitration, Representative John D. Dingell, Ranking Minority Member of the House Committee on Commerce, and Representative Edward J. Markey asked GAO to update its 1992 review of securities arbitration programs. Specifically, they wanted to know (1) whether arbitration forums had implemented GAO's recommendations and assessed the effectiveness of the changes; (2) how investors fared in securities arbitration; and (3) the extent to which investors were paid their arbitration awards. Results in Brief The securities industry SROs-the National Association of Securities Dealers (NASD) and the New York Stock Exchange (NYSE)-implemented GAO's 1992 recommendations by giving arbitration participants a larger role in selecting arbitrators, periodically surveying arbitrators to verify background information, and improving arbitrator training. NASD and NYSE generally assessed arbitrator performance using methods such as participant evaluations, focus groups, and discussions with arbitration attorneys. They said that this participant feedback indicated that the program changes have improved investors' perceptions of arbitrator performance and the fairness of the arbitration process. GAO could not reach conclusions about the fairness of the arbitration process from case outcome statistics. Investors did not receive as high a percentage of favorable arbitration awards during any year from 1992 through 1998, 49 to 57 percent, as they had during the period of January 1989 through June 1990, 59 percent. The percentage of the amount claimed that was awarded also declined during this period, 46 to 57 percent, compared with the earlier period, 61 percent. However, an increase in the percentage of cases settled during this period, generally 50 to 60 percent of the total cases concluded, may have changed the mix of cases going to a final arbitration award decision. For some of these cases, broker-dealers may have chosen not to arbitrate because they reasonably expected to lose. For this reason, the declining win rate could indicate little or no change in the fairness of the arbitration process. GAO could not apply the same technique it used in 1992 to determine whether arbitration results implied anything about the fairness of SRO-sponsored arbitration because the caseloads at an independent forum, the American Arbitration Association (AAA), and at the courts were too small to make meaningful comparisons. More importantly, however, GAO's survey of investors who received awards in 1998 found that a number of broker-dealers that had left the securities industry often did not pay arbitration awards rendered against them. GAO's survey found that 49 percent of the awards rendered in 1998 were not paid at all and an additional 12 percent were only partially paid. GAO estimated that the amount of unpaid awards was about $129 million, or 80 percent of the $161 million awarded to investors during 1998. About $13 million, or 8 percent of the unpaid awards, were still being disputed in court through such actions as a motion to vacate or modify the award. Nearly all of the unpaid awards were from cases decided in NASD's arbitration forum. When investors complained, NASD took action to suspend nonpaying broker-dealers and had success in recovering awards, but it did not monitor the payment of arbitration awards. In addition, most of the unpaid awards resulted from broker-dealers that were no longer in business. NASD is considering changes to its processes to reduce costs and increase options for investors, but these changes still may not address the problem of defunct brokers not paying awards. Ultimately, recovering losses caused by undercapitalized, financially irresponsible, or unscrupulous broker-dealers is difficult-if not impossible-for investors. Educating investors about the risks of doing business with such broker- dealers could help them avoid situations that may result in unpaid awards. In addition, investors' arbitration attorneys have suggested alternative approaches to address the problem of unpaid awards. This report contains recommendations to the Chairman of the U.S. Securities and Exchange Commission (SEC) regarding regulatory actions needed to address the problem of unpaid awards. SEC, NYSE, and NASD generally agreed with the report's recommendations, but expressed concern about certain aspects of GAO's analysis. They also suggested several technical changes to the report, which GAO has included where appropriate. Background Arbitration, an alternative to suing in court, is a process that uses a neutral third party to resolve differences between two parties in controversy. Most broker-dealers require their customers to sign predispute arbitration agreements that require disputes to be resolved through SRO-sponsored arbitration. Investors can initiate arbitration proceedings by filing claims with SRO-sponsored arbitration forums and paying filing fees. The forums then provide both parties with a list of potential arbitrators from which they can select who will arbitrate their dispute. The arbitrators' decisions are final and can only be appealed to the courts for narrowly defined reasons, such as arbitrator misconduct or bias. Arbitration awards are to be paid within 30 days from the date of the award, unless a party seeks judicial review of the award. The SROs, primarily NASD and NYSE, administer and oversee securities arbitration programs. SEC is responsible for overseeing the SROs' operations to ensure that they comply with securities laws and rules, including those that pertain to arbitration. Principal Findings SROs Have Made Program Changes Consistent With GAO Recommendations NASD and NYSE have initiated new arbitration processes that respond to GAO's 1992 recommendations. For example, NASD no longer suggests arbitrators for particular cases but produces a computer-generated list by rotation of eligible arbitrators from which the parties are to select. NYSE has also provided the parties more autonomy in selecting their arbitrators, and both SROs periodically check arbitrators' backgrounds and require them to take training. The SROs made these changes to improve the arbitration process and participants' confidence in arbitration. To assess arbitrator performance, the SROs obtained participant feedback on various aspects of their arbitration experience through evaluation forms. The SROs used the results of these evaluations to identify the need for training programs, to counsel arbitrators about deficiencies or problems, and to evaluate arbitrator performance. However, many participants did not complete the evaluation forms. SRO officials said that the forms that were completed indicated improvement in participants' perceptions of arbitrator quality. In addition, these officials said that they periodically discussed arbitration processes and specific program changes with various participants, including focus groups, and conducted limited surveys of participants to assess the effects of specific changes. These officials also said the results of these processes have shown that perceptions of arbitrator performance and the fairness of the arbitration process have improved. Arbitration Award Decisions Did Not Favor Investors as Often as They Did Before 1992 The percentage of cases in which arbitrators' decisions favored investors has declined from the 59 percent GAO found for cases decided from January 1989 through June 1990. The percentage favoring investors averaged about 51 percent for the period of 1992 through 1996, then increased to a high of 57 percent in 1998. The amount of awards made to investors as a percentage of what they claimed also declined during 1992 through 1998, from the 61 percent that GAO previously found to an average of about 51 percent of the amount claimed. However, the extent to which arbitration cases were settled before reaching an award decision increased. From 1989 to 1992 less than 50 percent of cases were settled, while for 1993 to 1998, the cases settled ranged between 50 to 60 percent. This increase in the percentage of arbitration cases settled may, in part, explain the decline in the percentage of awards that favored investors. GAO could not use the results of securities disputes at independent forums to gauge the fairness of SRO-sponsored arbitration as it had in 1992. AAA's securities caseload declined significantly after 1991. The number of cases AAA decided averaged only about 35 a year from 1992 through 1998, compared with nearly 250 cases in the January 1989 through June 1990 period that GAO included in its 1992 report. The percentage of favorable awards for investors at AAA declined from 1992 through 1998 from 88 to 50 percent, but these percentages are not statistically significant because of the small number of cases. GAO also could not compare arbitration results to those achieved by investors who sued brokers in court because so few court cases were decided. GAO found 121 securities-related disputes between individual investors and their broker- dealers at 5 federal district courts. Of the 121 disputes, the courts decided 15 cases (12 percent), the parties settled 22 cases (18 percent), and the courts dismissed 85 cases (70 percent) for such reasons as being remanded to arbitration or transferred to another district. In the 15 decided cases, which were not enough to be statistically significant, investors won 11 cases (73 percent). Information was not available to determine the percentage of claimed amounts awarded. Many Investors Were Not Paid Their Awards GAO estimated that about 500 NASD awards to investors in 1998 either were unpaid or were partially paid. GAO developed its estimates by surveying a random probability sample of 247 of the 845 investors who received monetary awards in cases decided in 1998. Nearly all of the nonpayments involved NASD-decided cases. NASD did not have procedures to monitor whether awards were paid, but it did follow up when investors complained. In 1998, investors filed 142 such complaints. These complaints were the only source of information NASD had on unpaid or partially paid awards. When it received written complaints, NASD sent letters to the nonpaying brokers or broker-dealers threatening to remove them from the securities industry. Although this did not always help investors recover their awards, in total, NASD suspensions or the threat of suspension resulted in about 40 percent of these unpaid awards being paid or otherwise settled to the satisfaction of the investors. Timely information about whether arbitration awards have been paid could help NASD better assist investors who have not been paid awards. Most of the unpaid awards resulted from broker- dealers that were no longer in business. NASD had no procedures to address this problem, but is considering changes to its arbitration program to help investors who have disputes with these broker- dealers. These changes include notifying the investors of their broker-dealers' status when the investors file an arbitration claim, limiting the ability of a defunct broker-dealer to enforce a predispute arbitration agreement against a customer, and establishing a rule to streamline the arbitration process when the broker-dealer is defunct and fails to appear. Such actions may help investors obtain more timely judgments against defunct broker-dealers, but these actions do not help them avoid broker-dealers that might not pay. SEC and the SROs have extensive investor-education programs, but these have not included data on the extent of award nonpayment. Publicizing this information, along with related investor-education information, might better focus investor attention on the possibility of unpaid arbitration awards. Encouraging investors to use the Central Registration Depository to more thoroughly evaluate the background of broker-dealers and individual brokers that they intend to do business with might help investors better make these important decisions. Some attorneys who have represented investors in securities arbitration have proposed alternative methods to better ensure payment of awards. These alternative methods include having the Securities Investor Protection Corporation (SIPC) cover unpaid awards, establishing a separate SRO- sponsored fund to pay awards, or increasing broker funds that are available to pay awards through additional capital or bonding requirements.2 GAO did not assess the feasibility of these proposals but obtained the views of SEC; NASD; and affected organization officials, including SIPC officials and those who represent arbitration attorneys and broker-dealers. Recommendations GAO recommends that the Chairman, SEC, require NASD to adopt procedures for monitoring the payment of arbitration awards. Such procedures should include requesting the parties in an arbitration to notify NASD, by the end of the 30- day payment period, about the payment status of any monetary award, so NASD can begin timely suspension proceedings against nonpaying broker- dealers, as appropriate. GAO recommends that the Chairman require NASD to develop procedures addressing the problem of unpaid awards caused by failed broker-dealers to help reduce costs and increase options for investors, such as the changes NASD is considering. GAO also recommends that the Chairman work with the SROs to (1) develop and publicize information to focus investor attention on the possibility of unpaid arbitration awards and (2) encourage investors to more thoroughly evaluate the backgrounds of broker-dealers and individual brokers with whom they intend to do business. Lastly, GAO recommends that the Chairman periodically examine the extent of nonpayment of SRO arbitration awards to determine the effectiveness of actions taken to improve the payment of awards. To the extent unpaid awards remain a problem, the Chairman should establish a process to assess the feasibility of alternative approaches to addressing this problem. Agency Comments NASD Regulation, Inc., the NASD subsidiary with primary responsibility for its arbitration program through the Office of Dispute Resolution; NYSE; and SEC provided written comments on a draft of this report, which are reprinted in appendixes IV, V, and VI. These organizations generally agreed that the draft report revealed a potentially serious problem with the nonpayment of arbitration awards by broker-dealers and individual brokers that had left the industry and agreed with GAO's recommendations to address this problem. However, they expressed specific concerns about some of the approaches GAO used and some of the results of its analysis that are discussed in detail in chapter 5. _______________________________ 1 Securities Arbitration: How Investors Fare (GAO/GGD-92-74, May 11, 1992). 2 SIPC is a nonprofit membership corporation of broker-dealers, which protects customers of failed broker-dealers against loss of cash and securities up to statutorily defined limits. Chapter 1 Introduction Page 17GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Arbitration is used to resolve disputes between broker-dealers or individual brokers and their public customers (called investors). As an alternative to suing in court, arbitration uses a neutral third party to resolve differences between two parties in controversy. The arbitration process in the securities industry has generally occurred in self-regulatory organization (SRO)- sponsored forums that are overseen by the Securities and Exchange Commission (SEC). The number of arbitrations in these forums has averaged about 6,400 cases a year since 1992, after more investors became involved in the markets and as broker-dealers required them to arbitrate disputes in SRO-sponsored forums. In addition, the Supreme Court ruled in 1987 that investors who had signed predispute arbitration clauses could be compelled to resolve their disputes with broker-dealers by arbitration.1 Before 1987, arbitration caseloads were fewer than 3,000 a year, but these caseloads increased in number between 1987 and 1992. How Arbitration Is Designed to Work Customers' experience with arbitration generally begins when broker-dealers require them, before opening an account, to sign a contract that includes a predispute arbitration clause. If a dispute subsequently arises between the customer and the broker-dealer, the customer can file an arbitration claim with the forum indicated in the predispute agreement and with any SRO of which the broker-dealer is a member. In a customer-initiated arbitration case, the customer files a statement of claim with the designated SRO-sponsored arbitration forum, along with a filing fee that is based on the amount in controversy. The customer must also submit a refundable deposit to cover the cost of the first prehearing or hearing session. The forum's director of arbitration serves the statement of claim on the broker-dealer or individual broker (called respondents) against whom the claim has been brought. The respondent has from 20 to 45 days, depending on the forum used, to answer the claim with any defenses and related claims. A single arbitrator can be used to resolve claims under $25,000, solely on the basis of the parties' claims, when a hearing is not requested. Unless otherwise requested by a party, a single arbitrator can also be used to decide claims that are greater than $25,000 but less than $50,000. For all other cases, unless otherwise agreed to by the parties, a panel of three arbitrators is appointed to hear the dispute. After the filing process, the director of arbitration provides the parties with a list of potential arbitrators, most of whom are public rather than industry arbitrators, to hear the dispute. The parties indicate their preference and may challenge specific arbitrators on the list. Once the panel of arbitrators has been selected, the panel conducts hearings that may last a day or more depending on the complexity of the case. Arbitrators are to render their decisions after the presentation of the evidence at the hearings. They are not required to provide a reason or a written opinion when they make an award decision. The statement usually provides the amount awarded and any other nonmonetary relief. The award is final and is only subject to court review for narrowly defined reasons, such as arbitrator partiality, fraud, or disregard of the law. However, a court can confirm an arbitration award to establish a judgment against the party owing the award. Unless a party files a motion to vacate or modify the award, awards are to be paid within 30 days of the award date. If the award is not paid within 30 days, interest will accrue at the legal rate or the rate specified by the arbitrators in the award. SROs Administer and SEC Oversees Securities Arbitration Securities arbitration proceedings are administered through arbitration forums sponsored by various SROs, such as the National Association of Securities Dealers (NASD) and the New York Stock Exchange (NYSE). SEC is responsible for overseeing the SROs' operations to ensure that they comply with securities laws and rules, including those that pertain to arbitration. The SROs are not-for-profit organizations that have primary regulatory responsibility under the Securities Exchange Act of 1934 to adopt and enforce standards of conduct for their members. The SRO arbitration forums are funded from several sources, such as claim filing fees, charges for the use of facilities, per case member surcharges and case processing fees, and subsidies from SRO general revenues. SEC oversees the SROs' rulemaking regarding arbitration and inspects SRO arbitration programs for compliance with the securities laws and SRO rules. The American Arbitration Association (AAA) also administers securities arbitration proceedings. AAA is an independent, not-for-profit organization that handles dispute resolution for several industries and is not regulated by SEC. AAA is funded fully from case processing fees. Parties in AAA arbitration cases also must pay the arbitrators' fee and other costs. Objectives, Scope, and Methodology Because of their continuing concerns about how investors fare in SRO-sponsored arbitration forums, and whether investors are paid their awards, Representative John D. Dingell, Ranking Minority Member, House Committee on Commerce, and Representative Edward J. Markey asked us to update our 1992 report on securities arbitration programs.2 Our objectives were to determine (1) whether securities arbitration forums had implemented our 1992 recommendations and assessed the effectiveness of the changes, (2) how investors fared in securities arbitration award decisions, and (3) the extent to which investors received the amounts awarded by arbitration panels. To obtain information on the changes to arbitration programs that NYSE and NASD made to address our 1992 recommendations and any assessments of those changes, we interviewed SEC and SRO arbitration forum officials. We also analyzed the revised arbitrator application forms, including arbitrator profile forms, and other background information the forums use to select individuals for the arbitrator pool. We interviewed forum officials about their verification of the information submitted. We also reviewed the arbitrators' manual and schedules of courses available and attended a course given to chairpersons of arbitration panels. In addition, we reviewed arbitration participant evaluation forms used by NASD and NYSE and their analyses of evaluation results. To obtain information on the use of predispute arbitration clauses, we sent a questionnaire to the top 10 broker-dealers by numbers of retail sales representatives listed in the Securities Industry Association's (SIA) Securities Industry Yearbook 1998-1999 and 2 broker-dealers that provide on-line brokerage services. Nine of the 12, including both on-line broker-dealers, responded to our questionnaire. To determine how investors fared in arbitration award decisions, we used various statistical techniques to analyze arbitration awards for cases involving investors in disputes with securities broker-dealers. We obtained information on securities arbitration awards from a database maintained by the Securities Arbitration Commentator, Inc., (SAC) of Maplewood, NJ. This database includes information on arbitration awards decided at all SRO-sponsored forums and AAA, and the database contained the only comprehensive, consistent data available without reviewing case files at all of the forums. However, because this database is limited to awards, we were unable to provide a detailed analysis of investors' claims that were settled or withdrawn. We performed tests to assess the reliability of the SAC database. Our methodologies and statistical procedures are discussed in detail in appendix I. To determine how investors fared in court, we reviewed court case records for five selected eastern U.S. districts-the District of Columbia, Maryland, northern Illinois, southern New York, and eastern Virginia. We chose these districts because of their proximity to major securities markets and convenient locations. To determine the extent to which arbitration awards were not paid, we surveyed a random sample of 247 awards from the 845 monetary awards to investors that we identified in the SAC database as being decided in 1998. Eighty-five percent of the sampled awardees responded to our survey. Survey estimates made in this report, which are projected to the entire population of 1998 awards, are subject to sampling error. Unless otherwise noted, all percentage estimates from the survey have sampling errors of plus or minus 5 percentage points or less, and all total dollar estimates have sampling errors of plus or minus 10 percent or less. Our survey methodology is discussed in detail in appendix II. Our survey questionnaire is shown in appendix III. We obtained additional information on unpaid 1998 awards from NASD records of written complaints and records related to its processing of those complaints. To obtain information on the status of broker-dealers identified as not paying an award, we sent NASD a list of those broker-dealers, and its staff determined the broker-dealers' status. To obtain information and views on possible solutions to the problem of unpaid awards, we interviewed regulatory officials and officials of the Securities Investor Protection Corporation (SIPC) and organizations that represent arbitration attorneys and broker-dealers. We requested comments on a draft of this report from the Chairmen of SEC, NASD, and NYSE. Their comments are discussed in chapter 5. Our work was conducted in Washington, D.C., and vicinity; New York, NY, and vicinity; Boca Raton, FL; Baltimore, MD; Chicago, IL; and San Francisco, CA, between October 1998 and May 2000. Our work was performed in accordance with generally accepted government auditing standards. _______________________________ 1 Shearson/American Express, Inc., v. McMahon, 482 U.S. 220 (1987). 2 Securities Arbitration: How Investors Fare (GAO/GGD-92-74, May 11, 1992). Chapter 2 SROs Have Made Changes to Improve Their Arbitration Programs Page 22GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Since our 1992 report, SEC has worked with NASD and NYSE to strengthen the arbitration process.1 These SROs have made several changes to improve their arbitration processes, many of which respond to our recommendations. To assess the effectiveness of program changes, the SROs sought participants' views on their satisfaction with arbitrator performance and the effectiveness of program changes. NASD and NYSE officials told us that they believed the overall process has improved and is fair for investors. SEC Has Worked With SROs to Improve Securities Arbitration Since 1992, SEC has worked with the SROs to strengthen the securities arbitration process through its rule reviews and inspections. SEC officials told us that one of the most important rule changes they recently approved involved NASD's arbitrator selection process, which is discussed in the next section. In addition, SEC has completed two arbitration program inspections each at NASD and NYSE since 1992. SEC is currently conducting an inspection of NASD's program and plans to begin an inspection of NYSE's program in 2000. Since 1992, SEC also has inspected arbitration programs at other SROs, such as the Chicago Board Options Exchange and the Pacific Exchange. According to SEC officials, SEC reviews SRO administration and processing of arbitration cases in each inspection. In addition, they said SEC examines SRO arbitrator pools, focusing on SRO's recruitment, training, and evaluation of arbitrators. In each inspection, SEC advises the SRO of any deficiencies found and recommends that the SRO implement remedial measures to correct the deficiencies. SEC also follows up with the SRO to ensure that the SRO implements SEC's recommendations. SEC staff further noted that one of the purposes of their inspections is to evaluate the impact of any changes in the arbitration process. SROs Have Made Changes to Improve Their Arbitration Processes NASD and NYSE officials told us that they have made several changes to improve their arbitration processes since our 1992 report. They said the changes address each of the report's recommendations, which were that the SROs should develop standards for selecting arbitrators, verify information submitted by arbitrators, and establish specific training for arbitrators. NYSE and NASD have also sponsored a symposium and task force, respectively, to address issues relating to the arbitration process. NYSE's Symposium on Arbitration in 1994 and NASD's Arbitration Policy Task Force in 1996 addressed the adequacy of SRO information on arbitrators and the arbitrators' experience. Both groups also identified needed changes in the arbitration selection process and the training of arbitrators. NASD Made Changes to Improve Its Program NASD officials told us that they have made significant changes to their arbitration program in response to our recommendations addressing arbitrator qualification, selection, and training. For example, NASD now requires 5 to 8 years of professional or practical experience for applicants to become NASD arbitrators. NASD also surveyed arbitrators in 1992 and again at the end of 1998 to review and verify the accuracy of information on their background. In addition, NASD officials said that they check background information when arbitrators are selected to decide disputes. NASD requires the arbitrators to sign an oath or affirmation indicating that the information on their background is accurate. NASD officials told us that one of NASD's most significant changes has been the change to its rules that provides for selection of arbitrators from a list. Under the previous rules, NASD staff provided the parties with a list of three arbitrators. The parties had one peremptory challenge and unlimited challenges for cause to eliminate particular arbitrators. The new rules provide for a list selection process that gives the parties a greater role in choosing who will decide their cases. Under the new process, NASD supplies a list of up to 15 names that are selected by computerized rotation of the arbitrator roster. The parties can strike anyone from the list and rank the remaining arbitrators according to their preferences. If the parties cannot agree, they are assigned the next available arbitrator on the computerized list to fill any remaining vacancies. NASD also revised the list of arbitrators, eliminating names for various reasons, such as unsatisfactory evaluations in previous arbitration cases, failure to complete new training requirements, lack of interest, or conflicts of interest that would prevent them from serving as independent arbitrators. In January 1993, NASD began requiring all new arbitrators to complete introductory training before becoming eligible to serve on a case. According to an NASD official, the training program has been refined in the past 5 years. One of the refinements, which began in March 1998, requires arbitrators to pass a test on arbitration procedures to become eligible to serve on a case. NASD also has taken action to separate its arbitration activities from its market and regulatory activities. In September 1999, SEC approved a new NASD subsidiary-NASD Dispute Resolution, Inc.-which is to be responsible for managing arbitration disputes. NASD reported that this change, which should be operational by the summer of 2000, would make the arbitration process more independent. An NASD official also said that NASD had made other changes to its arbitration program, including encouraging parties to try mediation to resolve disputes before going to arbitration, issuing a Discovery Guide to define which documents should be exchanged among the parties in the arbitration process, streamlining the claim process, and establishing a Web site on arbitration issues. NYSE Made Changes to Improve Its Program NYSE officials told us that NYSE, among other things, has established a new arbitrator profile for maintaining information on arbitrators' background and qualifications. NYSE also has developed procedures for verifying this information and established training standards for arbitrators. The profile serves as the initial disclosure of information about arbitrators to the parties, counsels, or witnesses in a particular case. The profile includes information on employment history, education, professional and arbitration experience, brokerage affiliation, and arbitration training. NYSE also requires potential arbitrators to have a minimum of 5 years experience in their chosen profession or to have two letters from members of their profession or the community that endorse arbitrator nominees and address their experience and character. NYSE officials said that this information assists them in selecting potential arbitrators. NYSE officials stated that NYSE continually checks arbitrators' qualifications. The officials noted that each time arbitrators are assigned a new case, NYSE asks the arbitrators to review, update, and sign their profile information. They said that NYSE also verifies this information annually for each arbitrator as well as checks arbitrators' disciplinary histories for past securities or criminal violations. As part of the arbitrator profile, NYSE is to disclose any disciplinary history that does not warrant the arbitrator's removal from the pool. Arbitrators who receive poor evaluations on more than one occasion may be removed from the pool. In addition, arbitrators who are suspended or barred by a regulatory organization or who fail to disclose their disciplinary history are also to be removed from the arbitrator pool. In addition, NYSE may (1) temporarily disqualify arbitrators if they are currently the subject of a complaint or investigation and (2) permanently disqualify arbitrators for failure to disclose material information or for being subject to a finding of fraud by a court or arbitration. NYSE is also trying a new process for selecting arbitrators. Under a pilot program, NYSE permits the parties in arbitration to choose from two options to select their arbitrators. NYSE either provides the parties with a list of 3 arbitrators, allowing specific opportunities to remove an arbitrator from the list, or the parties can request random lists of up to 15 arbitrators. NYSE officials told us that they have received favorable responses from the parties on the available alternatives. NYSE has also conducted arbitrator training seminars on procedures, issues, and ongoing developments concerning securities arbitration. It requires all new arbitrators to attend the arbitrator-training program and expects all arbitrators to periodically attend additional training. In addition, NYSE officials said that NYSE has an arbitration Web site that provides information to investors, including on-line access to all arbitration awards issued since 1992. These officials said NYSE also has pilot programs that encourage the use of mediation and more efficient case processing through administrative conferences with the arbitrators to resolve preliminary matters. SROs Seek Participants' Views to Assess Effects of Program Changes To assess the effects of their program changes, NASD and NYSE use participant evaluations to obtain investor perceptions about arbitrator performance on individual cases. Officials from these SROs said that their review of the evaluations has shown a high level of participant satisfaction with arbitrators' performances. These officials said that they also use other methods to evaluate the effects of program changes, such as focus groups and meetings with individual arbitration participants. Overall, these officials said that the arbitration process has improved and is fair for investors. The evaluation forms that the SROs use focus on arbitrators' skills and traits, such as display of professionalism and sensitivity during the proceedings, whether the case was conducted according to prescribed procedures, and whether arbitrators displayed knowledge of the securities industry. Forum officials said that they use these evaluations to develop training programs, counsel arbitrators about deficiencies or problems, and determine if certain arbitrators should continue to be on the list of arbitrators for the specific forum. NASD officials told us that getting participants to complete the evaluation forms has been extremely difficult for the forums over the years. The January 1996 Arbitration Policy Task Force Report2 recommended that NASD make a greater effort to improve participants' response rate. Shortly after the report was published, NASD developed a new evaluation form to elicit a higher participant response rate. An independent analysis of NASD's participant responses by professors of the Department of Social Sciences at the United States Military Academy at West Point found that for 2,037 cases closed by hearing between December 1, 1997, and April 1, 1999, the response rate varied from 10 to 20 percent. NASD officials told us that although they strive to have a greater degree of participation, West Point professors told these officials that the response rate provided a valid statistical sample. For example, the professors' analyses showed that, of the parties responding, 93 percent indicated that their cases were handled fairly and without bias. An NASD official said NASD plans to continue such analyses of the parties' evaluations to identify overall trends in responses. The participant evaluation forms have not included questions on process changes. SRO officials said that this use of the evaluations to assess program changes is inappropriate because investors rarely have more than one experience in arbitration and, therefore, have no point of reference from which to gauge the effect of specific changes to the program. SEC officials noted that the sole purpose of the evaluations is to assess arbitrator performance. The SROs have also used other methods to assess participant satisfaction and the effectiveness of program changes. During 1999, NASD conducted focus groups with customer and industry participants to obtain feedback and comments about its new list selection process. NASD also surveyed selected attorneys who were frequent participants in arbitration to assess the effectiveness of the arbitrator selection program. NYSE officials said that they also met regularly with participant attorneys to assess overall satisfaction with the NYSE program. SEC officials also noted that NASD and NYSE arbitration program officials, as well as those of other SROs, meet quarterly with representatives of both public investors and the industry through the Securities Industry Conference on Arbitration (SICA).3 These officials said that SICA has provided a useful forum for the SROs to assess the effectiveness of specific arbitration program changes. _______________________________ 1 Securities Arbitration: How Investors Fare (GAO/GGD-92-74, May 11, 1992). 2 Securities Arbitration Reform: Report of the Arbitration Policy Task Force to the Board of Governors National Association of Securities Dealers, Inc., January 1996. 3 SICA, formed in 1977, is a cooperative effort of representatives of the securities industry; the SROs; and the public to implement a uniform arbitration system, monitor that system, and change it as appropriate or required. SEC staff said they also attend SICA meetings. Chapter 3 Arbitration Awards Favoring Investors Declined, but More Cases Were Settled Page 32GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Our analysis of arbitration results from 1992 through 1998 showed that the percentage of cases decided in favor of investors had declined compared with cases decided before 1992. The awards investors received as a percentage of the total amount that they claimed had also declined. However, during this period, the percentage of cases settled without an arbitrator's decision has increased, which SEC officials said may explain the declines in awards favoring investors. We reviewed several factors that might affect award outcomes, such as representation by an attorney or size of claim. The factors that affected case outcomes were similar to those we identified in 1992. In 1992, we also compared the results of arbitration in SRO-sponsored forums with those in an independent forum, AAA, to provide an indication of the fairness of SRO-sponsored arbitration. However, AAA's securities caseload declined significantly after 1990, and although AAA awards to investors also declined from 1992 to 1998, its caseload is no longer large enough to provide a meaningful comparison. Also, as in 1992, we could not compare arbitration results to those achieved by investors in litigation because the processes were different and so few cases were decided in court. Without comparison to such benchmarks, statistics on case outcomes and settlements provide limited information about the fairness of the arbitration process. Investors Have Not Fared as Well in Award Decisions as They Did Before 1992, but More Cases Were Settled In 1992, we reported that 59 percent of investors received favorable decisions in cases arbitrated from January 1989 through June 1990. This percentage declined to an average of about 51 percent for cases decided from 1992 through 1996 and increased to 56 percent in 1997 and 57 percent in 1998. (See fig. 3.1.) Similarly, we reported in 1992 that investors who had favorable decisions received an average of 61 percent of the dollar amount they claimed. This percentage declined to an average of about 51 percent during 1992 through 1998, ranging from a low of 46 percent in 1994 to a high of 57 percent in 1997.1 (See fig. 3.1.) Figure 3.1: Percentage of Decisions Favoring Investors and Awards to Amounts Claimed, January 1989 Through June 1990 and January 1992 Through December 1998 Source: GAO analysis of arbitration award data. Regulatory officials attribute the decline in arbitration results favoring investors to a corresponding increase in the percentage of cases settled. These officials said the higher percentage of cases that are settled tends to reduce the percentage of arbitration award decisions in which investors might receive favorable awards because broker-dealers are more likely to try to settle cases that they think they might lose. We estimate that between 1992 and 1998, the percentage of cases settled ranged from 43 percent in 1992 to 50 percent or greater for the years 1993 through 1998, reaching a high of 60 percent in 1997. In 1992, we reported that settlements occurred in 44 percent of SRO cases and 33 percent of AAA cases. The NASD and NYSE arbitration forums decided most of the cases that were arbitrated during 1992 through 1998. For example, NASD's arbitrators decided 1,428 cases, or 92 percent, of the total 1,552 customer-initiated arbitration cases in 1998. NYSE decided 90 cases, or 6 percent of these cases. The other forums accounted for the remaining 2 percent of these cases.2 We could not determine the reasons for the differences in overall case outcomes from year to year. Each arbitration case has different participants and varying circumstances relating to the claims and, therefore, must be judged on its own merits. Attempting to evaluate these factors would have involved time-consuming detailed reviews of case files, extensive examination of the relationships among these variables, and subjective judgments about the merits of individual cases. Also, differences in the membership among SROs could affect the result, and the investment products involved could be more complex and risky from one year to the next. In addition, investors having similar claims could simply be asking for higher amounts from one year to the next, which could change the percentage of the claim awarded with no difference in the type of case or the amounts awarded. Factors Affecting Awards for Investors For arbitration cases decided from 1992 through 1998, we used a multivariate analysis to evaluate how certain factors affected investors' chances of winning an award and the size of the award relative to the amount claimed. This analysis allows simultaneous evaluation of the effects of several factors on a particular result and estimation of the effects of any one factor by controlling or holding constant all other factors. (See app. I.) Factors Affecting Arbitration Results Without adjusting for differences in membership, our multivariate analysis showed that when other factors are controlled, investors were 23 percent more likely to receive a favorable decision at NASD between 1992 and 1998 than at NYSE. Investors received favorable decisions in an average of about 54 percent of the cases at NASD during this time period as compared with 45 percent at NYSE. For the most recent year, as shown in table 3.1, the percentage of all cases decided in favor of investors in 1998 at NASD was 57 percent compared with 50 percent at NYSE. Of the 34 cases decided at the other forums, 15 favored investors. Table 3.1: Percentage of Cases Decided in Favor of Investors in 1998 Forum Percent Industrywidea 57% NASD 57 NYSE 50 aIncludes SRO-sponsored forums and AAA. Source: GAO analysis of arbitration award data. SEC and NYSE officials told us these differences could be attributed to differences in NASD and NYSE membership. Securities regulations require every broker-dealer that has public customers to register with, and be a member of, NASD. As a result, they said NASD members include the newer, less established or less capitalized broker- dealers that may be more likely to take actions or fail to exercise proper internal controls over actions that could cause them to lose arbitration decisions. NYSE members must be able to pay the cost of a seat on the exchange (one recently sold for about $2 million); therefore, these members generally include the larger, more established broker-dealers. When adjusted for the differences in membership at NASD and NYSE, our analysis showed an insignificant difference in the results of their arbitration decisions. We compared the results of cases for large broker-dealers that were members of both NASD and NYSE and had multiple awards in both forums. These awards accounted for over 31 percent of all investor-initiated arbitration cases decided from 1992 through 1998. At NASD, investors won 44 percent of the awards. At NYSE, they won 41 percent of the awards. NASD arbitrators awarded investors more than 50 percent of the amount they claimed 37 percent of the time, while NYSE awards were more than 50 percent of the amount claimed 33 percent of the time. In addition to looking at different forums' effect on cases decided by arbitration, we also looked at other factors to determine whether they could have affected case outcomes. These factors were (1) the type of disposition (i.e., customer-member, customer-employee, or small claim); (2) the year of arbitrator's decision; (3) attorney representation; (4) claim components, such as compensatory damages, punitive damages, or attorney fees; (5) filing of counterclaim; (6) whether the process included a hearing of the evidence or a review of written evidence; (7) the processing time; (8) the total claim amount; and (9) the number of hearings. Our analysis indicated that investors were 27 percent more likely to receive an award if an attorney represented them. In 1992, attorney representation did not affect whether investors received favorable decisions. The components of the claim also were correlated to case outcome. Investors filing claims that included more than compensatory damages, such as punitive damages, interest, or various fees, were more than twice as likely to receive an award. We did not address this factor in 1992. Cases that involved a higher number of hearing sessions were also more likely to be decided in favor of investors.3 Investors who had between 5 and 10 hearing sessions were 27 percent more likely to receive favorable decisions, and if they had 11 or more hearings, they were 113 percent more likely to receive a favorable decision than claims involving fewer than 5 hearings. In 1992, we reported that investors were about 40 percent more likely to receive an award if they had a hearing than investors whose cases were decided only after a review of written evidence. Whether a broker-dealer filed a counterclaim against an investor also affected case outcomes. A broker-dealer may file a counterclaim against the investor after receiving the investor's claim. Investors were 43 percent more likely to receive a favorable decision if the broker-dealer did not file a counterclaim. In 1992, counterclaims were not factors in case outcomes. Our analysis also shows that investors' chances of receiving a favorable decision in arbitration decreased from 1992 to 1995 but have increased since 1995. In 1998, investors were 33 percent more likely to receive a favorable decision than they were in 1995. Detailed results of our analysis are shown in appendix I. Factors Affecting the Amount Awarded Without adjusting for differences in membership, our multivariate analysis showed that investors who received a favorable decision at NASD were 45 percent more likely to receive an award in excess of 50 percent of their claim than investors who received favorable decisions in other forums. In 1992, we found that forum did not affect the amount of award. In addition, investors with attorney representation in cases occurring at any forum were 30 percent more likely to receive an award greater than 50 percent of the amount claimed. In 1992, investors represented by attorneys were 60 percent more likely to receive an award in excess of the average award amount. The size of the claim also affected award amounts and the odds of winning a sizable award varied by the year they were awarded. Claims involving $10,000 to $100,000 were only 60 percent as likely as claims under $10,000 to receive an award over 50 percent of the amount claimed. Claims involving more than $100,000 were only 27 percent as likely to receive an award over 50 percent of the amount claimed. We reported in 1992 that claims under $20,000 were nearly 4 times as likely as larger claims to result in an award greater than the average percentage of the amount claimed. By 1998, the odds of winning a greater than average award were 30 percent higher than in 1992 and 50 percent higher than in 1995. We could not determine from our analysis whether the amounts claimed were justified or excessive. Unless the investor requests a hearing or the arbitrator calls one, claims of $25,000 or less are decided by one arbitrator after a review of written evidence. Our analysis showed that investors who had arbitrators decide their cases on the basis of the written evidence were about 40 percent more likely to receive an award over 50 percent of the amount claimed than those that had hearings. Other Comparisons in Arbitration Case Processing In addition to the factors that could have affected case outcomes, we also analyzed differences in the size of investor claims, the processing times, and the types of claims initiated by investors among SRO-sponsored forums. For example, the median investors' claim industrywide in 1998 was $64,000, and the median claim was higher at NYSE than at NASD. Table 3.2 provides the range of claim amounts filed for the industry and for NYSE and NASD in 1998. Table 3.2: Size and Range of Investors' Claim Amounts in 1998 Forum Investors' claim amounts Low High Mean Median Industrywidea $200 $40,563,0 $320,00 $64,00 00 0 0 NASD 200 23,500,00 288,000 63,200 0 NYSE 700 7,500,000 384,000 75,000 aIncludes SRO-sponsored forums and AAA. Source: GAO analysis of arbitration award data. Table 3.3 shows the average time it took to decide a case at SRO-sponsored forums based either on reviews of written evidence or a hearing. Decisions made after a hearing took considerably more time than ones made only on the basis of the written evidence. Smaller claims can be decided solely on the basis of the written evidence. To determine the time it took, we used the dates from when the forum received an investor's claim to when the forum sent the arbitrators' decisions to the parties. Disputes decided at NASD took longer to resolve than those at NYSE. NASD officials told us that their cases took longer because of their larger caseload and the parties' increased involvement in the process. Table 3.3: Average Processing Time for Cases Decided in 1998 (Days) Forum Average processing time (days) Hearing Written submission Industrywidea 504 281 NASD 519 289 NYSE 311 134 aIncludes SRO-sponsored forums and AAA. Source: GAO analysis of arbitration award data. Investors attempting to recover compensatory damages was the primary reason that cases entered arbitration. An award of compensatory damages may include the party's actual dollar loss and any other damages, such as interest or lost profits. A party in arbitration can also claim punitive damages, or compensation in excess of actual damages, which are intended to punish wrongdoers. In addition, a party may claim attorney fees and other expenses of the arbitration process, such as forum fees. Investors claimed punitive damages in about 20 percent (317 cases) of the 1,552 total cases decided in 1998. This percentage was less than the 28 percent we reported for cases decided by SRO forums in 1992. Arbitrators awarded punitive damages in 107 cases, or about 34 percent of the 317 decided cases in which investors requested such damages. This was a significant increase from the 12 percent we reported in 1992. Investors claimed reimbursement for attorney fees in about 10 percent (148 cases) of the 1,552 total cases decided in 1998. This percentage had decreased considerably from the 30 percent we reported in 1992. Investors received attorney fees in 100, or 68 percent, of the 148 cases in which they claimed such fees. This percentage was considerably more than the 17 percent we reported in 1992. Arbitrators also awarded attorney fees in 44 cases in which investors had not requested such fees. We also obtained information on lost interest and other costs associated with the claim that we did not include in our 1992 report. Investors claimed these costs in 14 percent (211 cases) of the total cases decided in 1998. They received these costs in 160 cases, or 76 percent of the 211 decided cases in which they claimed these costs. Use of AAA Has Declined In 1992, we compared SRO-sponsored arbitration results to those at AAA, as an indication of the fairness of the arbitration process. We found no significant differences in case outcomes. We could not make the same comparison for this report because AAA's securities caseload declined significantly from what we reported in 1992 and, therefore, AAA's results no longer provided as meaningful a comparison to SRO-sponsored forums. Predispute Agreements Require the Use of SRO- Sponsored Forums Since our 1992 report, large broker-dealers' policies requiring predispute arbitration agreements have expanded to include cash accounts in addition to margin and option accounts. The nine broker-dealer firms that replied to our survey required individual investors to agree to resolve their disputes through SRO-sponsored arbitration as a condition of opening most types of accounts.4 The only exceptions were for investors who opened plain cash accounts. As shown in table 3.4, six of the nine broker-dealers told us that they require predispute agreements to open at least some retail cash accounts; the remaining three broker-dealers did not require such agreements. In 1992, we reported that eight of the nine large broker-dealers that responded to our survey did not require individual investors who opened retail cash accounts as of December 1, 1990, to sign agreements containing arbitration clauses. All nine of the broker-dealers that responded to our survey and had margin accounts said they required predispute agreements for customers who opened these accounts. For broker- dealers having option accounts, eight of nine said they required such agreements. Table 3.4: Firms Requiring Predispute Arbitration Clauses in Accounts for Individual Investors as of July 1999 Account types Predispute arbitration clause Accounts Required for Required Not that all for required did not use such some for accounts accounts such accounts Retail Plain Cash 0 4 2 3 Retail IRA Cash 0 8 1 0 Retail 401 K Cash 2 6 1 0 Retail Margin 0 9 0 0 Retail Options 1 8 0 0 Institutional 1 5 2 1 Source: GAO analysis of broker-dealer survey data. The large broker-dealers we surveyed did not offer their customers AAA as a choice to arbitrate a dispute. Some of the reasons the large firms gave for their policy of using SRO-sponsored forums were lower costs, a more timely process than at the independent forums, and arbitrators' knowledge of the securities industry. Few Results at AAA According to AAA officials, broker-dealer predispute agreements that require investors to use SRO-sponsored arbitration forums have caused AAA's securities-related caseload to decline significantly. In our 1992 report, we analyzed the results of 248 AAA investor-initiated, securities- related awards from the period of December 1989 through June 1990. As table 3.5 shows, from 1992 through 1998, AAA averaged only about 35 such awards a year, ranging from a low of 20 awards in 1998 to a high of 49 in 1996. Also, as shown in the table, investors received favorable decisions and large percentages of their claims often in the early 1990s and then both results declined through the years, until 1998 when investors received favorable decisions in 50 percent of the cases and received 13 percent of their claims. Table 3.5: AAA Investor-Initiated, Securities- Related Awards and Results for Investors, 1992 Through 1998 Year Percentage of Percentage Investor- favorable of claims initiated, decisions for awarded securities- investors related awards 1992 24 88% 79% 1993 25 80 72 1994 43 72 52 1995 46 63 56 1996 49 53 48 1997 35 54 48 1998 20 50 13 Source: GAO analysis of arbitration award data. AAA officials told us that they were still interested in deciding securities arbitration disputes. In January 2000, SICA started a 2-year cooperative pilot program to give brokerage customers the option to use non-SRO forums to arbitrate disputes. NASD, NYSE, and other SROs are to cooperate in the pilot program, and seven retail brokerages committed to participate. Under the program, customers that have qualified claims with one of the participating brokerages may have the option of having their dispute heard at a non- SRO forum designated by the brokerage. The two participating non-SRO forums are AAA and the Judicial Arbitration and Mediation Service. Few Securities Disputes Are Litigated, Most Are Dismissed As in 1992, the results of arbitration and court cases are not comparable because of the inherent differences in the processes5 and their respective outcomes and the small number of litigated cases. Most of the securities-related court cases we reviewed were dismissed. We identified 817 securities and commodities cases that were terminated (decided or dismissed by the court or settled by the parties before a court decision) between January 1997 and December 1998, at the 5 federal district courts we visited. The courts' information systems did not distinguish securities cases from commodity cases, but we reviewed the 817 cases and identified 121 to be securities- related disputes between individual investors and their broker-dealers. Of the 121 disputes, 15 (12 percent) were decided in court; 22 (18 percent) were settled by the parties before a court decision; and 85 (70 percent) were dismissed for various reasons, such as they were remanded to arbitration or were transferred to another district. The 15 cases the courts decided, 11 in favor of the investor, were not statistically significant. We could not determine what percentage of investor claims the courts awarded because the claims often were not quantified. For the 10 cases in which the claims were quantified, the investors were awarded (1) the full compensatory amount claimed in 7 cases and (2) the full punitive amount in 5 of the 7 cases in which punitive damages were requested. The average time to litigate the 15 cases was 930 days and the median time was 1,151 days. The average time to settle the 21 cases was 1,045 days; the median time was 644 days. The average time to litigate the 23 cases we reviewed for our 1992 report, was 744 days; the median time was 594 days. The settlement time averaged 510 days, and the median time was 365 days. _______________________________ 1 Our analysis did not calculate the percentages of claim amounts awarded separately for compensatory and punitive damages. 2 The other three forums were AAA, Chicago Board Options Exchange, and the Pacific Exchange. 3 Cases that required a large number of hearings may have many different characteristics that affect their outcomes from those that required fewer hearings. We did not have information on these characteristics, and they are not identified in our analysis. Therefore, our analysis should not be interpreted as implying that prolonging cases will positively affect their outcomes. 4 We surveyed the use of predispute arbitration agreements for the 10 largest broker-dealers by number of retail representatives and 2 additional brokers that provide on-line brokerage services. We sent questionnaires to the 12 broker-dealers and 9 responded. 5 For example, the amount of discovery allowed and rules of evidence are different. See appendix I of our 1992 report (GAO/GGD-92-74). Chapter 4 Many Arbitration Awards Were Not Paid Page 37GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards On the basis of our survey of investors who received arbitration awards during 1998, we estimated that 49 percent of the awards were not paid, and an additional 12 percent were partially paid.1 Our estimates showed that these investors did not receive nearly 80 percent of the $161 million that they were awarded.2 Nearly all of the unpaid awards involved arbitration cases decided in NASD's arbitration forum. Investors who complained to NASD had some success in collecting their awards when NASD initiated suspension proceedings against the nonpaying broker-dealers after receiving the complaints. However, NASD did not routinely monitor the payment of awards to ensure that it took timely action against the nonpaying broker-dealers. Such actions could provide some investors with a better chance to collect their awards. Better follow up on award payments, however, will not address the primary nonpayment problem, because most broker-dealers that failed to pay the awards were no longer in business. To the extent these broker-dealers may be insolvent, investors have little chance to recover their awards. NASD's arbitration program did not address this problem, but after discussing our preliminary findings, NASD suggested some program changes that might reduce costs and increase options for investors. In addition, some investor arbitration attorneys have suggested other approaches that might be considered to address the nonpayment issue. Over Half of Awardees Were Not Paid As table 4.1 shows, an estimated 61 percent (7 percentage points) of investors who won arbitration awards in 1998 either were not paid or received only partial payment. This percentage estimate rises to 64 percent (7 percentage points) if only NASD cases are considered. We developed these estimates from a survey of a random probability sample of 247 of the 845 investors who received monetary awards in arbitration cases decided in 1998. Table 4.1: Estimates of the Percentage of Monetary Customer Arbitration Awards Issued by All Forums and NASD in 1998 That Broker-Dealers or Individual Brokers Paid Nothing, Paid Partially, or Paid Fully Forum Number of Percentag Percentage monetary e paid paid Percenta customer nothing partially ge awards paid fully All forums 845 49%a 12% 39%a NASD 786 52a 12 36a aSampling error is 7 percentage points. Source: GAO analysis of awardee survey data. On the basis of responses from investors who reported total dollars unpaid, we estimated that unpaid awards amounted to $129 million, or 80 percent (8 percentage points), of the $161 million total awarded in 1998.3 The unpaid awards included an estimated $55 million ($8 million) of unpaid awards for punitive damages. Table 4.2 shows that, in general, larger awards were less likely to be paid than smaller awards. For example, only 5 percent of awards of $1.15 million and over were paid in full, compared with 44 percent (9 percentage points) of awards under $100,000. Also, only 17 percent (6 percentage points) of the total dollars owed in awards of $1.15 million and over were paid, compared with 43 percent (12 percentage points) of the total dollars owed for awards under $100,000. Table 4.2: Estimates of Percentage of Award Cases Paid in Full and Dollar Amounts Paid in 1998, by Size of Award Range of Percentage Total Percentage award of awards dollars of total amount Number of paid awarded dollars awards in full paid in part and in full $1.15 22 5% $82,281,6 17%a million and 00 over $277,000 up 73 20 24 to $1.15 39,350,60 million 0 $100,000 up 130 38b 21,858,90 37b to $277,000 0 Under 620 44c 17,318,87 43d $100,000 0 aSampling error is 6 percent. bSampling error is 17 percent. cSampling error is 9 percent. dSampling error is 12 percent. Source: GAO analysis of awardee survey data. Our survey allowed respondents to cite one or more reasons for nonpayment or partial payment of their awards, on the basis of their knowledge of the cases. The most frequently mentioned reasons were � the broker-dealer was out-of-business, estimated to be a reason in 53 percent (9 percentage points sampling error) of all late or unpaid awards; � the broker-dealer claimed to be financially unable to pay the award, estimated at 32 percent (8 percentage points); � an individual broker (associated person) owing part or all of the award could not be located, estimated at 28 percent (8 percentage points); or � the broker-dealer had filed for bankruptcy, estimated at 21 percent (7 percentage points). We also estimated that for 3 percent4 of the unpaid awards, the award was not paid because the award had been modified or vacated by a court, or such a measure was pending, which are legitimate reasons for nonpayment. In an estimated 8 percent5 of the cases, nonpayment or partial payment was in some way due to the occurrence of a postaward settlement, a compromised award (in which the awardees agreed to accept less than the full award), or an installment payment arrangement. Most survey respondents with unpaid or partially paid awards reported taking a variety of actions to collect their awards. Survey respondents said they complained to the forum in an estimated 71 percent (10 percentage points) of the cases in which they received an award; 67 percent (10 percentage points) said they complained to the broker-dealer, an individual broker, or their attorneys; 67 percent (11 percentage points) said they took "further legal action"; and 34 percent (11 percentage points) reported filing a complaint with SEC about the unpaid award. Forums in our sample, other than NASD, decided too few awards in 1998 to be projected with meaningful results. Of the 44 NYSE arbitration cases decided in 1998 in which investors received monetary awards, we sampled 10 and received 8 responses. Seven of the eight respondents said their awards were fully paid, and one said the principle amount of the award was paid but the respondent declined to pay the forum fee, which he was awarded.6 Of the 10 AAA awards in 1998, we sampled 4 and received responses from 3. Two of the three respondents said their awards were fully paid, and one said nothing was paid. We did not receive a response from the one Pacific Exchange awardee in our sample. The number of completely unpaid NASD awards was estimated to be 410 (60 percentage points), representing nearly all (99.7 percent) of the 411 (60 percentage points) unpaid awards in all forums. Partially paid NASD awards numbered about 95 (37 percentage points), comprising most (94 percent) of the 101 (39 percentage points) partially paid awards for all forums. After removing non-NASD awards from our analysis, we estimate that 52 percent (7 percentage points) of the 786 NASD customer awards issued in 1998 were completely unpaid, 12 percent were partially paid, and 36 percent (7 percentage points) were paid in full (see table 4.1). NASD Did Not Routinely Determine If Arbitration Awards Were Paid Our estimates indicate that 504 (60 percentage points) NASD awards made in 1998 were either not paid or partially paid, but NASD records showed that it only received written complaints of nonpayment in 142 cases.7 These complaints were the only source of information NASD had on unpaid or partially paid awards. When it received a complaint, NASD initiated suspension proceedings against nonpaying broker-dealers and often had success in compelling them to pay awards. After an NASD arbitration panel renders an award, NASD is to send copies of the award to the claimant and the respondent, including a cover letter noting that all monetary awards are to be paid within 30 days of receipt. NASD may suspend or cancel a broker-dealer's membership or an individual broker's registration if they are the respondents in an arbitration award and fail to comply with the award or an arbitration settlement agreement. When NASD receives a complaint that a member firm or individual broker failed to pay an award or settled amount, it sends a warning letter indicating that the respondent's membership/registration will be suspended unless one or more of five conditions had been met. To avoid being suspended, the respondent (or his/her attorney) must send NASD documentary evidence showing that � the award has been paid, � the parties have agreed to installments or have otherwise agreed to settle the matter, � the award has been modified or vacated by a court, � an action to modify or vacate the award is pending in a court, or � a bankruptcy filing is pending or a bankruptcy court has discharged the award. NASD revocation procedures also allow the respondent to request a hearing on the matter to consider whether (1) the respondent was given notification of the award, (2) the respondent satisfied the award, and (3) a valid reason exists for the respondent's failure to comply with the award. NASD sent warning letters to all of its member broker-dealers and individual brokers who were involved in the 142 investor complaint cases about nonpayment of 1998 awards. In 49 cases, the NASD member paid the award, agreed to an installment plan, or otherwise settled the matter without being suspended. Members in 14 cases sought a court order to vacate the award; 5 filed for bankruptcy, and 9 requested a hearing (3 of which also filed a motion to vacate). The parties settled one of the nine cases before the hearing. The hearing decisions for the other eight cases resulted in two award payments, three suspensions, and two dismissals because the members had terminated their NASD membership and were no longer subject to NASD actions. The final case was dismissed because the member filed a motion to vacate the award. NASD suspended the other 68 members because they failed to respond to the warning letter, 4 of these members eventually paid the award and were reinstated. In total, NASD action resulted in about 40 percent of the awards being paid or otherwise settled to the satisfaction of the investors. NASD officials told us they did not attempt to enforce the payment of awards without receiving a complaint of nonpayment. They said that attempts to enforce payment might give the appearance of favoring one side in arbitration over the other. They also said that NASD members are liable for paying interest on unpaid awards, which encourages prompt payment. Although remaining neutral is important to help maintain the credibility of an SRO-sponsored forum, ensuring that investors are paid arbitration awards is an important part of maintaining their confidence in arbitration. Some proactive actions can be taken that would not affect NASD's position. For example, as we previously discussed, NASD sends letters to both parties notifying them of the award and the 30-day payment requirement, then takes no further action unless investors complain. However, the large difference in our estimates of the number of investors who reported that they did not get paid and the number of complaints on which NASD took action, may indicate that many unpaid investors do not complain to NASD or that NASD responds only to written formal complaints. Thus, nonpaying brokers may continue to operate with no action being taken. Taking action to monitor the status of award payments and identify nonpaying brokers as soon as payment is due not only could benefit investors but also could help preserve their confidence in the arbitration process and thus benefit the entire securities industry. NASD could determine the status of award payment by requiring both parties to notify it when they have paid an award or received payment, at least by the end of the 30-day payment period. Data developed through this process could help NASD timely identify nonpaying broker-dealers or individual brokers and limit their ability to continue to operate without paying awards. This could be especially important because our survey results showed that a few broker-dealers were responsible for several nonpayments. For example, 10 broker-dealers were each responsible for 3 or more unpaid awards, which in sum accounted for 62 unpaid awards-about 25 percent of the 247 awards in our sample. NASD Is Considering Changes to Address Award Nonpayment by Broker-Dealers That Are No Longer in Business The primary cause of unpaid arbitration awards was broker-dealers and individual brokers who were no longer in business and had left the securities industry. NASD's arbitration program had no procedures to help investors deal with these failed broker-dealers. As a result, investors had to go through the arbitration process to obtain a judgment against these broker-dealers whether or not they were still in business. This would have been both expensive and time-consuming to the investors. NASD is considering making several program changes that might help investors better deal with failed broker-dealers. Our survey respondents identified 65 broker- dealers that had failed to pay at least 1 award. NASD officials reported to us that 13 of these broker-dealers were still active, while 52 went out of business after they had been expelled, suspended, terminated, or canceled from NASD membership or liquidated. According to NASD information, 25 of these broker-dealers were known to conduct business in microcap stocks. All 25 of these microcap broker-dealers were among the 52 identified as no longer in business. Our survey also identified 15 individual brokers that respondents indicated had not paid awards. We did not have enough information on them to accurately identify the individuals and determine their status. As the statistics show, investors who filed claims against broker-dealers that were no longer in business had little likelihood of getting paid their awards. SEC officials noted that to help recover all or part of their awards, these investors may confirm their arbitration awards in court. Doing this, the officials said, gives investors the ability to convert awards into judgments, which they then can try to enforce against any assets of the failed broker-dealer or individual broker that can be located. NASD officials said they are considering ideas for several program changes to reduce costs and increase options for investors seeking to establish a claim and judgment against defunct brokers. First, they would propose to NASD's Board and SEC a rule amendment that a respondent broker- dealer that has been terminated, suspended, or barred from NASD, or that is otherwise defunct, cannot enforce a predispute agreement against a customer in an NASD forum. Second, they also would propose a rule amendment to provide streamlined default proceedings for cases in which the terminated or defunct broker-dealer does not appear, but the investor affirmatively elects to pursue arbitration. Third, they would propose to advise investors that make arbitration claims of the status of the broker-dealer (such as terminated, out of business, or expelled) so these investors could better evaluate whether to continue with arbitration or to proceed in some other forum, such as in court. Although the changes NASD is considering may not ultimately result in investors receiving their arbitration award payments from broker-dealers that are no longer in business, they could reduce the costs of the process and increase the options available for investors. Other Potential Solutions to the Problem of Unpaid Awards The broker-dealer or individual broker that an investor chooses might ultimately influence the investors' chances of recovering an arbitration award because many nonpayments were caused by failed broker-dealers. Regulatory officials told us that failed broker-dealers were often undercapitalized, financially irresponsible, or unscrupulous. SEC and the SROs have education programs to help investors understand the potential risks of dealing with such broker- dealers, but these programs have not included data on the extent of award nonpayment. Publicizing such data might better focus investor attention on the possibility of unpaid awards. In addition, encouraging investors to use the Central Registration Depository8 to more thoroughly evaluate the background of a broker-dealer or individual broker could help investors who are considering opening brokerage accounts to avoid doing business with potentially troublesome brokers. Some attorneys who have represented investors in securities arbitration have proposed other methods to ensure that investors are paid their arbitration awards. Their proposals include � providing SIPC coverage of unpaid awards, � establishing a separate SRO-sponsored fund for unpaid awards, and � increasing the availability of funds from broker-dealers or individual brokers to pay awards by raising net capital requirements or requiring additional bonding or insurance to cover malpractice claims. We did not assess the feasibility of these proposals but obtained the views of SEC, NASD, and affected organization officials, including SIPC officials and officials of organizations that represent arbitration attorneys and broker- dealers. The first proposal would provide for payment of unpaid arbitration awards from SIPC funds. Congress enacted the Securities Investor Protection Act in 1970, which established SIPC, after the failure of many broker-dealers raised fears of a run on solvent broker-dealers. The statute's purpose is to encourage investors to leave their securities with broker-dealers by protecting the firms' custodial function. The act's provisions are designed to effect an expeditious return of all customer securities and cash held at a failed broker-dealer. SIPC officials told us that changing SIPC's mission to include coverage of all unpaid arbitration awards would require amending the Securities Investor Protection Act. The officials said coverage of unpaid awards would increase SIPC's caseload and require it to become larger and increase its need for resources. For example, SIPC, which protects only against broker-dealer insolvency in certain situations, had 29 employees at the end of 1998. In contrast, the United Kingdom's Investor Compensation Scheme (ICS) had over 115 employees as of March 1999. ICS is similar in function to SIPC in that it protects investors against insolvency, but ICS provides broader coverage of investor losses due to fraud. SIPC officials also noted that many arbitration awards are granted by default when a broker-dealer, that has gone out of business, fails to contest a claim. Thus, they said, the validity of the claim was never tested. Officials of SIA told us that expanding SIPC coverage to include all unpaid arbitration awards, if funded by assessments on broker-dealers, could increase costs for broker-dealers and investors. They also said that expanded coverage might encourage frivolous arbitration claims and reduce incentives for investors to carefully choose their brokers and investments. SEC officials agreed that expanded coverage could create moral hazards. In addition, SEC officials noted that paying possibly $129 million (our estimate of the amount of unpaid claims in 1998) a year to cover such claims would quickly exhaust the SIPC fund ($1.1 billion), which is intended as a reserve against the failure of a large broker-dealer. The fund was built up over the last 3 decades by assessing broker-dealers a percentage of their yearly gross revenues. SEC officials pointed out that the majority of this funding came from the handful of broker-dealers that consistently have yearly revenues far in excess of the vast majority of all active broker-dealers. The task of replenishing the fund would again fall largely on the shoulders of these few broker-dealers, which SEC said are well capitalized, have adequate reserves to cover arbitration judgments, and are not delinquent in paying such awards. We are reviewing SIPC operations in detail at the request of the Ranking Minority Member, House Committee on Commerce, and plan to report the results in a separate report. The second proposal made by the arbitration attorneys was to establish a separate SRO- sponsored fund to cover the compensatory damages part (actual investor losses) of unpaid arbitration awards. Arbitration attorneys suggested that such a fund could be financed from either one or a combination of (1) interest from the SIPC fund, (2) a charge on investor transactions, (3) fees on broker-dealers and individual brokers, and (4) funds obtained from NASD money penalties. These attorneys suggested that the fund be limited to the payment of investor claims of compensatory damages after an award is deemed "uncollectable," that is, when the liable party has left the securities industry. SEC and industry officials said that establishing such a fund could pose the same disadvantages as expanding SIPC coverage, including increased costs for broker-dealers and investors and the discouragement of broker and investor diligence. SEC officials also emphasized that the SIPC fund stands as a reserve against the failure of a large broker-dealer, and that the interest from the fund pays for SIPC's operations and the costs of small firm liquidations, of which there are approximately seven new proceedings each year. These officials said that diverting the interest to satisfy unpaid arbitration awards would deplete the fund's principal and could create the funding problems previously discussed. SEC officials noted that the level of the SIPC fund dropped from about $1.196 billion to $1.129 billion during the 1999 calendar year. The third proposal was to increase the funds available to brokers to pay arbitration awards by increasing net capital requirements. Arbitration attorneys suggested that additional broker-dealer capital could be set aside in escrow for a period of time, such as 2 years after their membership is terminated, so that these funds would be available to pay awards when the broker-dealer leaves the industry. SEC officials said that the purpose of the net capital rule is to require broker-dealers to maintain sufficient liquid assets to be able to self-liquidate in an orderly manner. This benefits investors by allowing insolvent broker- dealers to remain operating long enough to transfer customer assets out of the firm. These officials said the rule relies on accounting principles and, therefore, is not equipped to create reserves against the potential for adverse arbitration awards that might arise in a broker- dealer's future. SEC officials said that to establish reserves against the type of arbitration awards that go unpaid (such as awards that are based on claims not covered by the Securities Investor Protection Act, and which bankrupt a firm or cause it to close) would necessitate a substantial increase in the minimum capital requirements. These officials noted that investors obtaining arbitration awards would be general creditors of the failed broker-dealer. Therefore, absent an amendment to the U.S. Bankruptcy Code, the reserves would need to be large enough to cover all creditor claims for this proposal to provide meaningful assistance to investors. SEC officials stated that, in their view, a sizable increase in the net capital requirements would force many small broker-dealers out of the industry, and unduly penalize those broker-dealers operating in a responsible manner. Moreover, the costs of maintaining this additional capital could eventually be passed on to investors, at least in part. SEC officials also noted that effectively barring small broker-dealers from entering the securities business by raising net capital requirements could hurt investors by limiting their choice of broker-dealers. Industry officials said that they did not know how much net capital requirements would have to increase to cover investor claims. In addition, SEC officials pointed out that, under the net capital rule, broker-dealers must immediately book a liability after receiving an adverse arbitration award. Therefore, to remain in business, the broker-dealer must maintain sufficient capital to cover the amount of the award. Furthermore, these officials said that under Generally Accepted Accounting Principles, a broker-dealer must record a liability for a pending arbitration claim if it is likely the broker-dealer will lose and if the amount of the pending award is reasonably certain. They said this accounting principle requires broker-dealers to evaluate all pending claims and determine whether the claims need to be booked as liabilities for net capital purposes. SEC officials stated that they would determine whether this evaluation process should be made more transparent to regulators by requiring thorough documentation. Other methods suggested to increase the availability of funds from broker-dealers to provide for payment of awards were to place additional bond requirements or have broker- dealers and individual brokers carry explicit insurance to protect against malpractice claims. Industry officials said these methods also could raise costs on broker-dealers industrywide and ultimately on investors. Industry officials were uncertain how much insurance would be needed and whether insurers would be willing to underwrite the coverage without limitation on liability. NASD Rule 3020 already requires that member broker- dealers maintain a blanket fidelity bond to protect against various losses, including fraudulent trading. Coverage varies from $25,000 to $500,000 depending on the broker-dealers' net capital requirement. However, the bond provides only first-party coverage, meaning only the broker- dealer could file a claim. Also, the required coverage, especially for small broker-dealers, might not be enough to pay for multiple awards exceeding $25,000. _______________________________ 1 Our estimates are based on survey responses of claimants or their representatives regarding the status of award payment from a representative random probability sample of awards favoring investors decided in 1998. We did not follow up to validate the accuracy of the survey responses. Survey estimates also are subject to sampling error. Unless otherwise noted, all estimates of percentages have sampling errors of plus or minus 5 percentage points or less, and all estimates of total numbers (e.g., of awards or dollars) have sampling errors of plus or minus 10 percent or less of those total values. This percentage estimate (49 percent) has a sampling error of 7 percentage points. 2 In 3 percent of these cases, survey responses indicated that the award had been modified or vacated by a court or such action was pending. If we exclude all of these awards from the sample because they legitimately may no longer be owed to the investors, our estimate of the unpaid dollar amount of the awards decreases to 72 percent. Our other estimates do not vary significantly when these cases are excluded. 3 When the 3 percent of unpaid awards for which survey respondents said the award was modified or vacated by a court or pending in court are excluded, the dollar amount of unpaid awards falls to $116 million, or 72 percent. Other estimates do not vary significantly when these cases are excluded. 4 The sampling error for this small estimate results in a 95-percent confidence interval around the value ranging from 1 to 10 percent. 5 The confidence interval ranges from 3 to 17 percent. 6 NYSE officials said that a review of the award indicates that the forum fee, which was $15, was not part of the award to the investor. 7 Our survey respondents said that they complained to the forum in an estimated 71 percent of the cases, or about 360 awards. This number vastly exceeds the 142 cases in which NASD records contained a complaint letter about a 1998 unpaid award. Similarly, SEC officials told us that they received far fewer complaints than the estimated 34 percent reported by our sample respondents. We do not know why the survey results and reported complaints received differed. 8 The Central Registration Depository is a database that NASD maintains containing employment and disciplinary histories of individual brokers as well as disciplinary actions taken against member broker-dealer firms. Chapter 5 Conclusions and Recommendations Page 46GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Conclusions NASD and NYSE have made changes to their arbitration programs that are consistent with our 1992 recommendations and intended to improve the fairness of the arbitration process. These changes appear reasonable, and NASD and NYSE evaluations have shown that the changes have improved both the process and investors' perceptions of its fairness. We could not determine whether the decline in the percentage of awards that favored investors indicated anything about the fairness of SRO- sponsored arbitration. An increase in the percentage of settled cases-some of which broker- dealers may have chosen not to arbitrate because they reasonably expected to lose-may have changed the mix of cases going to arbitration. For this reason, the declining win rates may not indicate a change in the fairness of the arbitration process. Unlike our 1992 report, we did not have sufficient cases from AAA to compare the results of an independent forum to the SRO-sponsored forums. Also, we did not find enough securities-related court cases to compare results with the SRO forums. However, the high dismissal rate of securities-related court cases may indicate that, in general, investors have not fared better in court. The securities industry, its regulators, and Congress should be concerned about the extent to which arbitration awards are unpaid. Regardless of how effective and fair the arbitration decision process may be, unpaid awards could negatively affect investors' confidence in arbitration. A timely NASD follow-up program on award payments would likely improve the chances that some investors would be paid their awards. In addition, investors would save time and money in establishing their claims if NASD implemented the types of actions it is considering. Ultimately, recovering losses caused by undercapitalized, financially irresponsible, or unscrupulous broker-dealers is difficult, if not impossible, for investors. However, data developed from monitoring award payments should help educate investors about the possibility of unpaid awards associated with doing business with these broker- dealers. In addition, encouraging investors who are considering opening brokerage accounts to use the Central Registration Depository to evaluate the background of a broker-dealer or individual broker could help them make better decisions. The arbitration attorneys' suggested award payment alternatives raise policy issues that warrant careful consideration and resolution before they could be effectively implemented. Recommendations We recommend that the Chairman, SEC, require NASD to adopt procedures for monitoring the payment of arbitration awards. Such procedures should include requesting the parties in an arbitration to notify NASD, by the end of the 30-day payment period, about the payment status of any monetary award, so NASD can begin timely suspension proceedings against nonpaying broker-dealers, as appropriate. We recommend that the Chairman require NASD to develop procedures addressing the problem of unpaid awards caused by failed broker-dealers to help reduce costs and increase options for investors, such as the changes NASD is considering. We recommend that the Chairman work with the SROs to (1) develop and publicize information to focus investor attention on the possibility of unpaid arbitration awards and (2) encourage investors to more thoroughly evaluate the backgrounds of broker- dealers and individual brokers with whom they intend to do business. Lastly, we recommend that the Chairman periodically examine the extent of nonpayment of SRO arbitration awards to determine the effectiveness of actions taken to improve the payment of awards. To the extent unpaid awards remain a problem, the Chairman should establish a process to assess the feasibility of alternative approaches to addressing the problem of unpaid awards. Agency Comments and Our Evaluation NASD Regulation, Inc. (NASDR), the NASD subsidiary with primary responsibility for its arbitration program through the Office of Dispute Resolution; NYSE; and SEC provided written comments on a draft of this report, which are reprinted in appendixes IV, V, and VI. These organizations generally agreed that our draft report revealed a potentially serious problem with the nonpayment of arbitration awards by broker- dealers and individual brokers that had left the industry, and they agreed with our recommendations to address this problem. However, these organizations expressed specific concerns about some of the analytic approaches we used and the results of our analysis. Also, NASD provided additional information on the actions it has taken to improve the overall efficiency and fairness of its arbitration program and to evaluate the effects of these changes. NASDR noted that the problem of unpaid awards does not indicate a problem with the NASD arbitration program but rather of bankrupt or defunct firms. It said that the same collection problems against these firms exist when investors take their claims to court or non-SRO arbitration forums. Nonetheless, NASDR agreed that appropriate measures need to be taken to encourage prompt payment of arbitrator awards and proposed several initiatives to address award nonpayment. If effectively implemented, these initiatives would comply with the intent of our recommendations. NASDR expressed concern about our conclusion that the statistics show arbitration awards favoring investors have declined since before 1992. NASDR cited data from SICA, both for all SROs and for NASD, that show the recent trend has been more favorable to investors with the composite figures from all SROs reaching higher levels in 1997 and 1998 than at any time in the previous 17 years. They said this trend continued in 1999 when investors won 61 percent of awards. The SICA data and the SAC database, which we used for our analysis, were developed from different sources, but produced similar results for the 1992 through 1998 period. However, the data from our previous report, which we developed from an extensive review of individual case files for cases decided over an 18-month period from January 1989 through June 1990, show a considerably higher percentage of favorable results for investors, 59 percent, than the yearly SICA data show for either 1989 or 1990, 53 percent for both years. We did not do any work during this review to attempt to explain these differences. NASDR also discussed its attempts to follow up on the unpaid awards we identified to determine whether current members were involved. This task was difficult because our survey results were confidential, and we could only provide NASDR with the names of broker-dealers and the number of unpaid awards attributed to each. NASDR's review of cases found only 21 that involved current member firms, and in 10 of those cases, NASDR reported that the firm had no obligation to pay the award. Because we only provided limited information, we do not know if the cases NASDR included in its review were the same cases that we surveyed. NASDR also questioned the methodology used to obtain the sample of awards we surveyed to obtain estimates of their payment. It was concerned that oversampling of the largest awards may have skewed the survey to include a greater percentage of awards against broker-dealers that are no longer in business. As described in appendix II, we sampled at a higher rate for large and medium awards than for small awards to provide more precise estimates for each stratum and across the entire population. However, in producing the estimates, we weighted responses to account statistically for all members of the population, including those that were not selected or did not respond to the survey. For example, all 95 of the largest awards were sampled, so they were each assigned a weight of 1 to represent only themselves. We assigned small awards larger weights to represent other small awards not sampled. Thus, the oversampling of large awards provided more precise estimates for those awards but did not overrepresent those large awards in our estimates for the entire population of awards. NYSE had no objections to our findings and recommendations but objected to our classification of one NYSE award as partially paid, solely on the basis of the response provided to our survey. NYSE validated that the award was fully paid, and we provided that information in our report. We did not use these data to project our sample results for NYSE. NYSE agreed with our finding that the decline in award decisions favoring investors could indicate little or no change in the fairness of the arbitration process. It further noted the information in the report that, in its view, supports the fairness of SRO arbitration. However, NYSE expressed concern about our analysis of factors affecting arbitration results with the SRO being one of the factors. NYSE noted that it is difficult to compare the results in arbitration unless similar claims and similar broker-dealers are used. We provided a separate analysis of awards against a common set of broker-dealers for NASD and NYSE but could not provide a similar distinction among claims. Any analysis of this sort is limited to the available data. We had information on several of the characteristics of cases and claimants, which we analyzed in a rigorous and systematic fashion. We acknowledge that more and better information might have altered our results. Unless the information we were missing is significantly associated with the characteristics we had available, our estimates, which were based on hundreds to thousands of award observations, provide unbiased estimates of the effects of those characteristics. SEC stated that the report provides useful data to confirm the message it has brought repeatedly to investors over the past 8 years: investors must investigate before they invest. SEC suggested that we could have provided more meaningful information about the fairness of SRO arbitration by comparing the procedures used by the SRO arbitration forums against independently developed standards for alternative dispute resolution. SEC said such an analysis would show that the SRO procedures measure favorably against the standards. Although measuring procedures against standards may provide a useful indicator of the fairness of the arbitration process, our requesters asked us to determine the outcomes of cases to assess how investors fared in securities arbitration award decisions. We could not comment on the fairness of the SRO arbitration process based on the statistics alone unless they could be measured against the outcomes of securities cases at an independent forum or the courts because these are the only other venues for resolving securities disputes. SEC also stated that our analysis of arbitration awards failed to distinguish between the amounts of compensatory damages claimed and awarded and punitive damages claimed and awarded. To the extent compensatory damages represent the actual losses of investors, a separate analysis of the awards for these damages may have provided useful information about whether investors recovered their actual losses through arbitration. We focused on the total amounts awarded because our objective was to determine how investors fared overall in arbitration award decisions rather than whether they recovered their actual losses. However, we include in chapter 3 data on the frequency with which punitive damages were claimed and awarded in 1998 that are comparable to data of our 1992 report. SEC also requested that our report provide greater detail on the payment of awards of less than $100,000 and from $100,000 to $277,000. It stated that this information would be useful for average investors who might decide to forego seeking restitution in the mistaken belief that arbitration awards are not paid. We now show these data in table 4.2. In addition, SEC asked that we clarify that broker- dealers that stay in business have a good payment record. Although, the information presented in our report suggests this may be true, we did not specifically analyze the payment records of these broker-dealers. Appendix I Methodology for Analyzing Arbitration Results Page 59GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards To determine how investors fared in securities arbitration, we analyzed arbitration award data. We used the award data to calculate rates (percentages) reflecting the extent to which investors won awards, damages investors' claimed were awarded, and various factors influenced arbitration outcomes. Arbitration Case Data The data we used to describe the characteristics of cases brought to arbitration and their outcomes were obtained from Securities Arbitration Commentator, Inc. (SAC), Maplewood, NJ. SAC is a commercial research firm that maintains a database of information from publicly available records on decided cases from all of the self-regulatory organization (SRO) arbitration forums and the American Arbitration Association (AAA). The SAC database contained information on arbitration awards that resulted from (1) customer (investor) claims of damages against SRO-member broker-dealers (called customer-member claims) or their individual registered representatives (called customer-employee claims) and (2) investors' small claims cases of less than a specified dollar amount (which only apply to customer-member claims), that SAC maintained as a separate award category, covering the period from January 1, 1992, through December 31, 1998. By definition, these data did not include records of cases that were settled or dismissed before a decision was reached. Estimates of the percentage of cases settled were from data provided by SICA. The 11,290 cases in the database included fields describing 113 variables such as the following: the name of the forum; the parties involved in the proceeding (customer-member, customer-employee, small claims, etc.); type of claim; amounts claimed; and amounts awarded. We analyzed the following cross tabulations to understand the data and plan the analysis: � the number of awards in favor of (won) and against (lost) investors, by year and by forum; � the difference between the amount awarded and the amount claimed, by year and by forum; � type of damages claimed,by year; � amount of damages claimed, by year and by forum; � type of award, by year and by forum; � amount of award, by year and by forum; and � processing time, by year and by forum. Our data reliability assessment included several steps. First, to assess the reliability of the data we selected a simple random sample of 350 arbitration award records in the SAC database and compared them to the corresponding hard-copy awards as issued by the forums that we obtained from SAC. This analysis determined the percentage of records that contained more than one inaccurate field and the percentage of inaccurate fields. The analysis showed that 4.0 percent of the sample records had errors in more than one field. Further analysis showed that the 95-percent confidence interval for the percentage of records with more than one error was 2.2 to 6.7 percent. The second analysis showed that 1.1 percent of the fields were incorrect. The 95-percent confidence interval bounds for the percentage of the fields that were incorrect ranged from 0.7 to 1.6 percent. Second, SAC also verified the completeness of awards in its database of closed awards through comparison to lists of closed awards at NASD Regulation, Inc., annual arbitration reports published by the Securities Industry Conference on Arbitration (SICA), and summary statistics of the National Association of Securities Dealers (NASD) arbitration program. Similar steps were taken to ensure receipt of all Municipal Securities Rulemaking Board and American Stock Exchange awards. AAA awards recorded in the SAC database were checked against AAA records by AAA. Awards rendered by the Pacific Exchange, The Chicago Board Options Exchange, Inc., and the Philadelphia Stock Exchange were routinely reviewed and compared with forum reports to spot any discrepancies with the SAC data. In 1998, SAC began the practice of checking with the New York Stock Exchange (NYSE) to determine that the NYSE awards contained in SAC's database represented all of the public awards rendered at the NYSE. However, because we could not be sure that any list or records of awards included all awards issued, we could not ensure that the SAC data included all awards issued from January 1992 through December 1998. To account for differences in membership among forums we calculated these percentages for a data set of awards involving major broker-dealers that had multiple awards in both NASD and NYSE. We identified the major broker-dealer cases as those involving the major broker-dealers in the industry in terms of registered retail sales representatives. We also limited our major broker- dealer analysis to those broker-dealers involved in multiple arbitration awards at both NYSE and NASD. On the basis of this definition, there were 3,502 major broker-dealer cases in the SAC database. These cases represented 31 percent of all of the cases in the database. To provide descriptive information on the extent to which investors won arbitration awards and the amount of damages claimed that was awarded, we calculated percentages for the amount of awards in which investors won and lost decisions and percentages for the amount claimed that was awarded. We analyzed these percentages by forum for all cases in the database. In addition, we analyzed won and lost percentages by forum for the award cases involving the selected major broker- dealers with multiple awards in both forums. For all cases in the databases (before selection of cases involving the major broker-dealers), percentages of investor-favorable decisions were 45 percent at NYSE, 54 percent at NASD, and 53 percent in all other forums. For only those cases selected as major broker-dealers, these percentages were 41 percent at NYSE, 44 percent at NASD, and 53 percent for all other forums. A significant number of investors received awards for less than half of the amount claimed across all forums. At NYSE, 60 percent of investor awards were for less than half of the amount claimed while at NASD, 50 percent of investor awards were for less than half of the amount claimed. The corresponding figure in all other forums was 40 percent. Similar results were found for the subset of cases that involved major broker-dealers with multiple awards in both NASD and NYSE. For example, at NYSE, 62 percent of investor awards were for less than half of the amount claimed while at NASD, 58 percent of investor awards were for less than half of the amount claimed. The corresponding figure in all other forums was 44 percent. Loglinear Analysis of Arbitration Outcomes We also used the SAC data to determine the factors that affected the outcome of securities arbitration in this time period. To characterize the features of arbitration cases that are associated with certain outcomes, we examined the influence of a number of factors in a multivariate model. We analyzed the data on arbitration cases in two steps. First, we investigated which characteristics of the claims of damages affected the likelihood that cases were won (decided in favor of the investor) rather than lost (decided against the investor). Second, we investigated, for claims of damages that were won and in which a monetary award was sought, which claim characteristics affected the likelihood that the amount awarded was more than 50 percent of the amount sought. The claim characteristics we considered included (1) the year of the claim (1992, 1993, 1994, 1995, 1996, 1997, or 1998); (2) the type of disposition (customer-member or customer-employee); (3) whether the claimant was represented by an attorney; (4) whether the claimant was seeking compensation only; (5) whether a counterclaim was made; (6) whether there was a hearing; (7) the time it took to process the claim (less than 6 months, 6 to 11 months, 12 to 24 months, or more than 24 months); (8) the forum in which the claim was brought (NASD, NYSE, or Other); (9) the claim amount sought ($0, $1 to $9,999, $10,000 to $100,000, or more than $100,000); and (10) the number of hearings (0, 1 to 4, 5 to 10, or more than 10).1 In these analyses, we determined the size and statistical significance of the effect that each of these claim characteristics had on whether cases were won, and on whether more than 50 percent of the amount claimed was awarded. We first looked at the effect of each characteristic ignoring every other, and then used multivariate logistic regression models to estimate the net effect of each of these characteristics, or the effects they had on these two outcomes when the associations between claim characteristics were controlled and all characteristics were considered simultaneously. Odds and odds ratios, which we describe below, were used to estimate the size of the effects of the different claim characteristics, and chi-square statistics and Wald statistics were used to determine whether they were statistically significant (i.e., large enough that they could not be assumed to be due to random fluctuations or chance). Our primary results are summarized in tables I.1 and I.2 below. The first table shows the effects of the various characteristics of arbitration cases on the likelihood that cases were won. The second table shows the effects of these same characteristics on the likelihood that cases that were won were decided for more than 50 percent of the total amount claimed (or sought) by the individual who filed the claim. Table I.1: Arbitration Cases Won and Lost, Odds on Winning Derived From Them, and Odds Ratios Indicating the Effects of Various Characteristics of Arbitration Cases on the Odds on Winning Variable/Category Arbitration Odds on Observ Predict result win ed ed odds odds ratios ratios Win Lose Total Type of claim filed Customer-Member 5,75 5,194 10,948 1.11 - - 4 Customer-Employee 191 151 342 1.26 1.14 0.85 Year 1992 973 846 1,819 1.15 - - 1993 782 754 1,536 1.04 0.90 0.88 1994 669 645 1,314 1.04 0.90 0.84 1995 771 808 1,579 0.95 0.83 0.76 1996 986 919 1,905 1.07 0.93 0.85 1997 886 699 1,585 1.27 1.10 1.02 1998 878 674 1,552 1.30 1.13 1.01 Representation Attorney 5,43 4,698 10,133 1.16 - - 5 No attorney 510 647 1,157 0.79 0.68 0.79 Claim components Compensation Plusa 2,24 1,004 3,246 2.23 - - 2 Compensation Only 3,40 3,999 7,404 0.85 0.38 0.43 5 Counterclaim made No 5,74 5,128 10,871 1.12 - - 3 Yes 202 217 419 0.93 0.83 0.70 Type of case Written 1,45 1,672 3,131 0.87 - - 9 Hearing 4,28 3,489 7,775 1.23 1.41 Omitted 6 Processing time < 6mos. 474 616 1,090 0.77 - - 6 mos. - 11 mos. 2,47 2,366 4,841 1.05 1.36 0.99b 5 12 mos. - 24 mos. 2,41 1,918 4,331 1.26 1.20 0.99 3 > 24 mos. 524 393 917 1.33 1.06 0.99 Forum NASD 5,08 4,411 9,493 1.15 - - 2 NYSE 509 621 1,130 0.82 0.71 0.81 Other 354 313 667 1.13 0.98 0.90 Total claim $0 374 354 728 1.06 - - $1-$9,999 1,64 1,944 3,584 0.84 0.79 0.75b 0 $10,000 -$100,000 1,78 1,674 3,462 1.07 1.01 0.78 8 >100,000 2,00 1,273 3,276 1.57 1.48 0.80 3 Number of hearings 0 1,45 1,672 3,131 0.87 - - 9 1-4 2,78 2,608 5,389 1.07 1.23 1.02 1 5-10 1,09 712 1,806 1.54 1.77 1.27 4 >10 411 168 579 2.45 2.79 2.12 Note: Odds ratios in the next to the last column are derived from the observed odds and indicate the effect of each claim characteristic when other characteristics are ignored. Odds ratios in the last column are from a multivariate model and indicate the effect of each claim characteristic when the effects of all other claim characteristics are statistically controlled. aCompensatory damages plus other damages, including punitive damages, attorney fees, and forum fees. bNot significant. Source: GAO analysis of arbitration award data. The odds on winning in table I.1 tell us how many claims were won for every claim that was lost. For the small number of customer-employee claims filed between 1992 and 1998, for example, the odds on winning (versus losing) were 191/151 = 1.26. This implies that 1.26 customer-employee claims were won for every one that was lost or, multiplying by 100, that 126 customer-employee claims were won for every 100 that were lost. For customer-member claims, the odds on winning were lower and equal to 1.11 (111 were won for every 100 that were lost). The odds ratio in the penultimate column of the table, which equals 1.26/1.11 or 1.14, tells us that the odds on winning were higher among customer-employee claims than among customer-member claims, by a factor of 1.14. We can also interpret this as meaning that the odds on winning were 14 percent higher among customer-employee claims than among customer- member claims. This odds ratio and the others in the same column indicate the effects each factor and the other factors had on award outcome when each factor is assessed while ignoring the others. The odds ratios in the final column of the table are more appropriate for assessing the effects these factors had on award outcome, since they are derived from statistical models that estimate the effects of each factor net of every other. Focusing on the predicted odds ratios in the final column of table I.1, our principal findings with respect to the odds on arbitration cases being won can be stated as follows: � When the effects of other factors were controlled, the type of claim filed, processing time, and amount claimed had no significant effects on whether arbitration cases were won or lost. � The year in which claims were filed had a significant effect. Between 1992 and 1995, the odds ratio declined gradually-in 1995 the odds on winning were only 3/4 of what they were in 1992 (or in 1992 they were about 32 percent higher than in 1995). Between 1995 and 1997-98, the odds on winning increased by 33 percent, so that, net of all other factors, the odds on winning ended up at the end of the period precisely where they were at the beginning of it. � Cases in which claimants were not represented by attorneys had lower odds on being won, by a factor of 0.79. Stated differently, the odds on winning were higher when attorneys represented claimants by a factor of 1.0/0.79 = 1.27, or by 27 percent. � Claims seeking only compensation were less than half as likely-or less likely by a factor of 0.43, to be precise-as other claims to be won. Alternatively, claims involving more than compensation were more than twice as likely (2.32 times as likely) to be won. � Claims in which no counterclaims were filed were 1.0/0.70 = 1.43 times as likely to be won, or 43 percent more likely. � NASD claims, not adjusted for differences in membership, were more likely than NYSE claims to be won, by a factor of 1.0/0.81 = 1.23, or by 23 percent. We found (in supplemental analyses not shown) this difference to be similar across all years. The small number of claims that were filed in other forums were somewhat less likely than NASD claims to be won and somewhat more likely than NYSE claims to be won, but they were not significantly different from either. � Claims that involved no hearings were not significantly different in terms of their odds on being won from those that involved between one and four hearings. Claims involving 5 to10 hearings were 27 percent more likely to be won, and claims involving 11 or more hearings were 113 percent more likely to be won than written claims. Table I.2 provides information pertaining to the second outcome we considered, involving whether cases that were won were decided for more than 50 percent of the amount claimed. Focusing on the odds ratios in the final column of table I.2, our principal findings with respect to the odds on arbitration cases that were won being decided for more than 50 percent of the total claimed (henceforth referred to as "the odds on winning sizable awards") can be stated as follows: � The size of the award was unaffected by whether more than compensation was sought, whether a counterclaim was filed, and processing time. � The type of claim filed had an immense effect on the odds on winning sizable awards, which was impossible to estimate with the data at hand. While slightly fewer than half of customer-member claims that were won received sizable awards, none of the customer-employee claims did. � The year in which claims were filed had a significant effect. Between 1992 and 1995, the odds on winning sizable awards dropped by roughly 10 percent. Between 1994-95 and 1998, the odds on winning a sizable award increased gradually. In 1998, the odds on winning a sizable award were roughly 30 percent higher than they had been in 1992 and 50 percent higher than in 1995. � Cases in which attorneys did not represent the claimants had lower odds on winning a sizable award, by a factor of 0.77. Stated differently, the odds on winning were higher when attorneys represented claimants by a factor of 1.0/0.77 = 1.30, or by 30 percent. � NASD claims, not adjusted for differences in membership, were more likely than NYSE claims to involve sizable awards, by a factor of 1.0/0.69 = 1.45, or by 45 percent. The small number of claims that were filed in other forums were somewhat less likely than NASD claims, and somewhat more likely than NYSE claims, to involve sizable awards, but they were not significantly different from either. � Larger claims were less likely than small ones to win sizable awards. Claims involving $10,000 to $100,000 were only roughly 60 percent as likely as smaller claims to win sizable awards, and claims involving more than $100,000 were roughly 27 percent as likely as smaller claims to win sizable awards. � Claims that involved no hearings were more likely than claims involving hearings to win sizable awards, though the number of hearings appeared to make little difference. In general, written claims were between 30 percent to 56 percent more likely than others to involve sizable awards (i.e., 1.0/0.77 = 1.30, or 30 percent; 1.0/0.64= 1.56, or 56 percent; and 1.0/0.69 = 1.45, or 45 percent). Table I.2: Arbitration Cases Won That Received Awards That Were More or Less Than 50 Percent of the Amount Claimed, and Odds Ratios Indicating the Effects of Various Claim Characteristics on Winning Sizable Awards Variable/Category Percentage of Total claim awarded Odds on Observed Predicte > 50 odds d odds percent ratios ratios < 50 > 50 perce percent nt Type of claim filed Customer-Member 2,821 2,565 5,386 0.91 - - Customer-Employee 183 0 183 0.00 Year 1992 506 397 903 0.78 - - 1993 407 323 730 0.79 1.01 1.06 1994 356 268 624 0.75 0.96 0.94 1995 387 335 722 0.87 1.12 0.86 1996 483 430 913 0.89 1.14 1.01 1997 428 412 840 0.96 1.23 1.20 1998 437 400 837 0.92 1.18 1.29 Representation Attorney 2,770 2,302 5,072 0.83 - - No attorney 234 263 497 1.12 1.35 0.77 Claim components Compensation Plusa 1,255 898 2,153 0.72 - - Compensation Only 1,744 1,659 3,403 0.95 1.32 0.94b Counterclaim made No 2,888 2,494 5,382 0.86 - - Yes 116 71 187 0.61 0.71 0.80b Type of case Written 487 934 1421 1.92 - - Hearing 2,454 1,546 4,000 0.63 0.33 Omitted Processing time < 6mos. 158 293 451 1.85 - - 6 mos. - 11 mos. 1,185 1,185 2,370 1.00 0.54 0.99b 12 mos. - 24 mos. 1,355 892 2,247 0.66 0.66 0.99 > 24 mos. 296 182 478 0.61 0.92 0.99 Forum NASD 2,553 2,254 4,807 0.88 - - NYSE 308 176 484 0.57 0.65 0.69 Other 143 135 278 0.94 1.07 0.93 Total claim $1-$9,999 553 1,087 1,640 1.97 - - $10,000 -$100,000 966 822 1,788 0.85 0.43 0.59 >$100,000 1,413 588 2,001 0.42 0.21 0.27 Number of hearings 0 487 934 1,421 1.92 - - 1-4 1,510 1,118 2,628 0.74 0.39 0.77 5-10 684 316 1,000 0.46 0.24 0.64 >10 260 112 372 0.43 0.22 0.69 Note 1: The analyses above pertain only to arbitration cases that sought monetary compensation and were won by the claimant. Note 2: Odds ratios in the next to the last column are derived from the observed odds and indicate the effect of each claim characteristic when other characteristics are ignored. Odds ratios in the last column are from a multivariate model and indicate the effect of each claim characteristic when the effects of all other claim characteristics are statistically controlled. aCompensatory damages plus other damages, including punitive damages, attorney fees, and forum fees. bNot significant. Source: GAO analysis of arbitration award data. _______________________________ 1 Preliminary analyses (not shown) revealed that the categories we used to represent the processing time, forum, claim amount, and number of hearings were suitable to capture the effects of these variables. That is, the large bulk of the variation in the outcomes we were looking at is between the categories we created, rather than within them. Appendix II Awardee Survey Methodology Page 62GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards To determine the extent to which monetary awards that arbitrators made were actually paid to investors, we surveyed a representative random probability sample of individual investors winning awards in 1998 cases. We used survey data to estimate the proportion of awards completely and partially unpaid, and the total dollar amount unpaid. Population and Sample Design Our target population consisted of all arbitration cases brought by individual investors against broker-dealers or their representatives that were closed in 1998 with a monetary award being granted. Using the SAC database to generate our sample frame, we initially identified an actual study population of 852 such award cases. We drew a stratified random probability sample of 250 awards. Three strata were created on the basis of the dollar amount of the award, and sample cases were selected at a relatively higher rate from the larger awards. See table II.1 for the allocation of the sample across the strata. With this statistically valid probability sample, each member of the study population had a nonzero probability of being included, and that probability could be computed for any member. The sample was designed to provide us with acceptably precise estimates of the proportion of awards paid and unpaid, and the total dollar amount of awards paid and unpaid, across the entire population. After drawing the sample, we discovered three pairs of duplicate elements, and an additional four pairs of duplicate entries in the sample frame derived from the SAC database that had not been randomly selected into the sample. This resulted in a final working sample size of 247 and a population total of 845. Administration of Survey Because the sample frame created from the SAC database did not contain mailing addresses and telephone numbers of awardees or their attorneys, we asked the SROs to provide that contact information for our sample of 247 awards. According to AAA officials, AAA's policy is not to release information that would identify claimants, so it was necessary to allow the company to mail our survey questionnaires directly to the claimants or their attorneys for the four AAA awards drawn into our sample. We chose to mail our questionnaires to the attorneys identified as the representatives for the awardees, rather than the awardees themselves, except in cases in which awardees represented themselves. We did this because the survey pretests suggested that attorneys were more likely to be able to provide the necessary information, and since they often worked for many clients, could sometimes answer for more than one of our sampled cases. For each sampled award, we prepared a questionnaire preprinted with the short caption (title) of the case, the names of parties involved, the award amount, and the date the arbitration decision was issued. The questionnaire requested that attorneys on that case (or the awardees themselves) indicate whether that particular award had been paid, how much of the payment had been received, whether payment was timely, reasons for nonpayment, and actions taken to collect the award. Appendix III contains a copy of the questionnaire. We mailed out the questionnaires on September 24, 1999, and asked AAA to mail questionnaires to their awardees simultaneously. On October 22, 1999, we sent a follow-up mailing with another copy of the questionnaire to the 116 sample elements from forums other than AAA who had not yet responded. On December 6, 1999, we began to make telephone calls to the 84 attorneys or awardees that had not yet responded, or for whom mailed questionnaires had earlier been returned as undeliverable. If we were successful in contacting nonrespondents by telephone, we conducted a telephone interview to complete as many of the items from the mail questionnaire as possible. On December 15, 1999, we ended fieldwork. Disposition of Sample We received 209 usable responses, which was an overall unit response rate of approximately 85 percent. Because not all respondents provided an answer to each question they were eligible to answer, the item response rates vary and are generally lower than the 85-percent unit response rate. Table II.1: Disposition of Sample Stratum Popula Sample Disposition Response tion size rate size Nonresponse Usable response Refus Undeliver All other Mail Telephone al able nonrespon response response se Large 22 22 0 0 1 15 6 95% awards Medium 73 73 2 1 6 41 23 88 awards Small 750 152 1 5 22 104 20 82 awards Total 845 247 3 6 29 160 49 85% Note: Large awards are those that are $1.15 million or greater, medium awards are $277,000 to less than $1.15 million, and small awards are under $277,000. Source: GAO analysis of arbitration award data. To produce the estimates from this survey, answers from each responding case were weighted in the analysis to account statistically for all the members of the population, including those that were not selected or did not respond to the survey. Survey Error and Data Quality Estimates from sample surveys are subject to a number of sources of error, which can be grouped into the following categories: coverage error, sampling error, nonresponse error, measurement error, and processing error. We took a number of steps to limit these errors. Surveys may be subject to coverage error, which occurs when the sampling frame does not fully represent the target population of interest. We could not ensure that the SAC data from which our frame was constructed included all awards (see app. I). We detected seven pairs of duplicate elements in our sample frame, which were removed before the survey was conducted. Sampling error exists because we followed a probability procedure that is based on random selections, and our sample is only one of a large number of samples that we might have drawn. Since each sample could have provided different estimates, we express our confidence in the precision of our particular sample's results as a 95-percent confidence interval (e.g., �7 percentage points). This is the interval that would contain the actual population value for 95- percent of the samples we could have drawn. As a result, we are 95-percent confident that each of the confidence intervals in this report will include the true values in the study population. Nonresponse error arises when surveys are unsuccessful in obtaining some or all information from eligible sample elements. To the extent that those not providing information would have provided significantly different information from those that did respond, bias from nonresponse can also result. Because the seriousness of this type of error is often proportional to the level of missing data, response rates are commonly used as indirect measures of nonresponse error and bias. We took steps to maximize response rates, such as multiple mailings and telephone follow-up to convert nonrespondents to respondents. Measurement errors are defined as differences between the reported and true values of the characteristics under study. Such errors can arise from the way questions are worded, differences in how questions are interpreted by respondents, deficiencies in the sources of information available to respondents, or intentional misreporting by respondents. To minimize such errors, we asked subject matter experts to review our questionnaire and pretested the questionnaire with several attorneys representing cases in our sample frame. We did not, however, verify the substance of answers given by awardees or their attorneys. Finally, surveys may be subject to processing error in data entry, processing, and analysis. We verified the accuracy of a small sample of keypunched records by comparing them to their corresponding questionnaires, and corrected errors found. Less than 1 percent of the data items we checked had random keypunch errors that would not have been corrected during data processing. Analysis programs were also independently verified. Appendix III Awardee Survey Questionnaire Page 65GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Appendix IV Comments From the NASD Regulation, Inc. Page 72GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Appendix V Comments From the New York Stock Exchange Page 74GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Appendix VI Comments From the Securities and Exchange Commission Page 77GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards Appendix VII GAO Contacts and Staff Acknowledgments Page 78GAO/GGD-00-115 Actions Needed to Address Pr oblem of Unpaid Awards GAO Contacts Thomas J. McCool (202) 512-8678 Michael A. Burnett (202) 512-8678 Acknowledgments In addition to the persons named above, Lamont Kincaid, Richard J. Hillman, Carl Ramirez, Barry Reed (Retired), Michelle A. Sager, Sidney H. Schwartz, Douglas M. Sloane, and David Tarosky made key contributions to this report. *** End of Document ***