Auditors and Analysts: An Analysis of the
Evidence and Reform Proposals in Light of the Enron Experience
I want to thank the Committee for inviting me to appear today. Because I realize that you are covering a broad range of issues and have only limited time to listen to any individual witness, I believe that my contribution will be the most useful if I focus on just two issues: (1) What powers, duties, and standards should Congress include in any legislation that establishes a self-regulatory body to oversee the auditing profession?; and (2) How should Congress respond to the evidence that conflicts of interest do bias the recommendations and research of securities analysts?
If we focus only on Enron, it cannot prove by itself that there is a crisis or that either auditors or securities analysts have been compromised by conflicts of interest. By itself, Enron is only an anecdote - - bizarre, vivid and tragic as it may be. But Enron does not stand alone. As I elaborated in detail in testimony before the Senate Commerce Committee on December 17, 2001 (and thus will not repeat at any length here), Enron is part of a pattern. As the liabilities faced by auditors declined in the 1990's and as the incentives auditors perceived to acquiesce in management's desire to manage earnings increased over the same period (because of the opportunities to earn highly lucrative consulting revenues), there has been an apparent erosion in the quality of financial reporting. Assertive as this conclusion may sound, a burgeoning literature exists on earnings management, which indicates that earnings management is conscious, widespread and tolerated by auditors within, at least, very wide limits.(1) Objective data also shows a decline in the reliability of published financial results. To give only the simplest quantitative measure, from 1997 to 2000, there were 1,080 earnings restatements by publicly held companies.(2) Most importantly, there has been a significant recent increase in the number of earnings restatements. Earnings restatements averaged 49 per year from 1990 to 1997, then increased to 91 in 1998, and soared to 150 in 1999 and 156 in 2000.(3) Put simply, this sudden spike in earnings restatements is neither coincidental nor temporary.
Worse yet, the accounting profession is conspicuous by its lack of any meaningful mechanism for internal self-discipline. This void contrasts starkly with the governance structure of the broker-dealer industry, where the National Association of Securities Dealers ("NASD") administers a vigorous and effective system of internal discipline. Because both brokers and auditors ultimately serve the same constituency - - i.e., investors - - this disparity is unjustifiable.
Put simply, American corporate governance depends at bottom on the credibility of the numbers. Only if financial data is accurate can our essentially private system of corporate governance operate effectively. Today, there is doubt about the reliability of reported financial data - - and also about the independence and objectivity of the two watchdogs who monitor and verity that data: namely, auditors and securities analysts.
What should Congress do about the crisis? While there is a case for raising the liabilities that auditors and analysts face, I am fully aware that many are skeptical of private enforcement of law through class and derivative actions. Essentially, this asks a third watchdog - - the plaintiff's attorney - - to monitor the failings of the first two (auditors and analysts), and plaintiff's attorneys may have their own misincentives. Also, it may still be too early to ask Congress to revisit the Private Securities Litigation Reform Act of 1995 (the "PSLRA"). Thus, both in my December appearance before the Senate Commerce Committee and again today, I am urging the Congress to give fuller consideration to public enforcement through the creation or strengthening of self-regulatory organizations ("SRO's"). An SRO already exists with jurisdiction over securities analysts (i.e., the NASD), but one needs to be created from whole cloth in the case of auditors. Thus, my comments will focus first on the creation of a new SRO for auditors and then how to strengthen the oversight of analysts.
II. An SRO for Auditors: Some Suggested Standards.
The governance of accounting is today fragmented and indeed Balkanized among (1) state boards of accountancy, (2) private bodies, of which there are essentially seven, and (3) the SEC, which has broad anti-fraud jurisdiction, but less certain authority under Rule 102(e) of its Rules of Practice.(4) Disciplinary authority is particularly divided within the profession. The Quality Control Inquiry Committee ("QCIC") of the SEC Practice Section of the American Institute of Certified Public Accountants ("AICPA") is delegated responsibility to investigate alleged audit failures involving SEC clients arising from litigation or regulatory investigations, but it is charged only with determining if there are deficiencies in the auditing firm's system of quality control. The Professional Ethics Executive Committee ("PEEC") of the AICPA is suppose to take individual cases on referral from the QCIC, but as a matter of "fairness" PEEC will automatically defer, at the subject firm's request, any investigation until all litigation or regulatory proceedings have been completed. In short, the investor's interest in purging corrupt or fraudulent auditors from the profession is subordinated to the firm's interest in settling litigation cheap, uninfluenced by any possible findings of ethical lapses.
Little in this system merits retention. Legislation is necessary to create a body that would have at least the same powers, duties and obligations as the NASD. In truth, however, the legislation that created the NASD in 1938 (the Maloney Act) is not an ideal model, given its general lack of specific guidance. Rather, model legislation should have the following elements:
1. Rule-making power. The SRO should be specifically authorized to (1) address and prohibit conflicts of interest and other deficiencies that might jeopardize either auditor independence or the public's confidence in the accuracy and reliability of published financial statements, and (2) establish mandatory procedures, including procedures for the retention of accountants by publicly-held companies and for the interaction and relationship between the accountants and audit committees. This is a broad standard - - and deliberately so. It could authorize the SRO to require that auditors be retained and/or fired by the audit committee and not by the company's management. In addition, the SRO should be authorized to affirmatively mandate the adoption and use of new or improved quality control systems, as they from time to time become accepted.
2. Mandatory Membership. All outside auditors preparing or certifying the financial statements of publicly-held companies or of companies conducting registered public offerings would be required to be members in good standing, and suspension or ouster from the SRO would render an auditor unable to certify the financial statements of such companies.
3. SEC Supervision. SEC approval of the initial registration of such an SRO and
of all amendments to its rules would be mandated, just as in the case of the NASD. The SEC would also have authority to amend the SRO's rules in compliance with a statutory "public interest" standard. Finally, the SEC should have authority to sanction, fine, or suspend the SRO and to remove or suspend its officers or directors for cause.
4. Enforcement Powers. The SRO should have the same authority to impose financial penalties or to suspend or disbar an auditor from membership, or to suspend, disbar, fine or censure any associated professional. Such fines and penalties should not require proof of fraud, but only a demonstration of negligent or unethical conduct. Subpoena authority should also be conferred, and a failure to cooperate or provide evidence should be grounds for discipline or dismissal.
5. Duties of Supervisory Personnel. A common response of organizations caught in a scandal or a criminal transaction is to blame everything on a "rogue" employee. Yet, such "rogues" are often responding to winks and nods from above (real or perceived) or to an organizational culture that encourages risk-taking (Enron is again symptomatic). The federal securities laws impose duties on supervisory personnel in brokerage firms to monitor their employees, and a parallel standard should apply to supervisory personnel in auditing firms.
6. Governance. The SRO should have at least a supermajority (say, 66 2/3%) of "public" members, who are not present or recently past employees or associated persons of the auditing industry.
7. Prompt Enforcement. The practice now followed by PEEC of deferring all disciplinary investigations until civil litigation and regulatory investigations have been resolved is self-defeating and unacceptable. It might, however, be possible to render the findings and disciplinary measures taken by the SRO inadmissable in private civil litigation.
III. Securities Analysts
A. What Do We Know About Analyst Objectivity?
A number of studies have sought to assess the impact of conflicts of interest upon the objectivity of securities analyst recommendations. Additional evidence was also recently collected at hearings held in June 2001 by the Subcommittee on Capital Markets, Insurance and Government-Sponsored Enterprises of the House Financial Services Committee. This data is probably more germane, and merits greater reliance, than the well-known statistic that an alleged 100:1 ratio exists between the "buy" recommendations and "sell" recommendations made by securities analysts. Although the actual ratio may be somewhat less extreme than 100:1,(5) the real problem with this statistic is that it is not necessarily the product of conflicts of interest. That is, analysts employed by brokerage firms (as all "sell-side" analysts are) have a natural incentive to encourage purchase or sale transactions. For this purpose, "buy" recommendations are more useful than "sell" recommendation, because all clients can buy a stock, but only existing holders can sell as a practical matter.
Other data better illustrates the impact of conflicts of interest on analysts. Among the most salient findings from recent research are the following:
1. Conflict of Interests. Several studies find that "independent" analysts (i.e., analysts not associated with the underwriter for a particular issuer) behave differently than analysts who are so associated with the issuer's underwriter. For example, Roni Michaely and Kent Womack find that the long-run performance of firms recommended by analysts who are associated with an underwriter was significantly worse than the performance of firms recommended by independent securities analysts.(6) They further find that stock prices of firms recommended by analysts associated with lead underwriters fall on average in the 30 days before a recommendation is issued, while the stock prices of firms recommended by analysts not so associated with underwriters rose on average over the same period. Finally, the mean long-run performance of buy recommendations made by analysts on non-clients is more positive than the performance of recommendations made on clients - - at least for 12 out of 14 brokerage firms.
Still another study by CFO Magazine reports that analysts who work for full-service investment banking firms have 6% higher earnings forecasts and close to 25% more buy recommendations than do analysts at firms without such ties.(7) Similarly, using a sample of 2,400 seasoned equity offerings between 1989 and 1994, Lin and McNichols find that lead and co-underwriter analysts' growth forecasts and particularly their recommendations are significantly more favorable than those made by unaffiliated analysts.(8)
2. Pressure and Retaliation. In self-reporting studies, securities analysts report that they are frequently pressured to make positive buy recommendations or at least to temper negative opinions.(9) 61% of analysts responding to one survey reported personal experience with threats of retaliation from issuer management.(10) Similarly, former Acting SEC Chairman Laura Unger noted in a recent speech that a survey of 300 chief financial officers found that 20% of surveyed CFO's acknowledged withholding business from brokerage firms whose analysts issued unfavorable research.(11) This is a phenomenon that is almost certain to be underreported.
This data should not be overread. It does not prove that securities research or analyst recommendations are valueness or hopelessly biased, but it does tend to confirm what one would intuitively expect: namely, conflicts of interest count, and conflicted analysts behave differently than unaffiliated or "independent" analysts.
B. The Regulatory Response
In light of public criticism regarding securities analysts and their conflicts of interest, the National Association of Securities Dealers ("NASD") proposed Rule 2711 ("Research Analysts and Research Reports") in early February, 2002.(12) Proposed Rule 2711 is lengthy, complex and has not yet been adopted. Nonetheless, because its adoption in some form seems likely, a brief analysis of its contents seems useful as an introduction to what further steps Congress should consider.
Basically, Rule 2711 does seven important things:
(1) It places restrictions on the investment banking department's relationship with the "research" or securities analyst division of an integrated broker-dealer firms;
(2) It restricts the pre-publication review of analyst research reports by the subject company and investment banking personnel;
(3) It prohibits bonus or salary compensation to a research analyst based upon a specific investment banking services transaction;
(4) It prohibits broker-dealers from promising favorable research or ratings as consideration or an inducement for the receipt of business or compensation;
(5) It extends the "quiet period" during which the broker-dealer may not publish research reports regarding a company in an IPO for which the firm is acting as a manager or co-manager for 40 calendar days from the date of the offering;
(6) It restricts analysts ability to acquire securities from a company prior to an IPO or to purchase or sell for a defined period before or after the publication of research report or a change in a rating or price target; and
(7) It requires extensive disclosure by an analyst of certain stock holdings or compensation or other conflict of interest relationships.
All of these prohibitions are subject to substantial exceptions and/or qualifications, and it is debatable whether some can be effectively monitored. Only time and experience with proposed Rule 2711 can tell us whether its exceptions will overwhelm the rule. Nonetheless, Rule 2711 represents a serious and commendable effort to police the conflicts of interest that exist within broker-dealer firms that both underwrite securities and provide securities research and recommendations. In this light, the most important question is: what else can or should Congress do? Are these topics or areas that Rule 2711 has not addressed that Congress should address? These are considered below:
C. Congressional Options
The overriding policy question is whether conflicts of interest relating to securities research should be prohibited or only policed. As I will suggest below, this question is not easily answered, because there are costs and imperfections with both options:
1. Radical Reform: Divorce Investment Banking From Securities Research. Congress could do what it essentially did a half century ago in the Glass-Steagall Act:(13) namely, prohibit investment banking firms that underwrite securities from engaging in a specified activity (here, providing securities research to all, or at least certain, customers). Arguably, this is what Congress and the SEC have already proposed to do with respect to the accounting profession: i.e., separate the auditing and consulting roles performed by accountants. Here, the conflict might be thought to be even more serious because the empirical evidence does suggest that the advice given by conflicted analysts is different from the advice given by independent analysts.
But this divestiture remedy is here even more problematic than in the case of the original Glass-Steagall Act. Put simply, securities research is not a self-sufficient line of business that exist on a free-standing basis. To be sure, there are a limited number of "independent" securities research boutiques (Sanford C. Bernstein & Co is probably the best known and most often cited example) that do not do underwriting, but still survive very well. Yet, this is a niche market, catering to institutional investors. Since May 1, 1975 ("Mayday") when the old system of fixed commissions was ended and brokerage commissions became competitively determined, commission have shrunk to a razor-thin margin that will not support the costs of securities research. Instead, securities research (i.e., the salaries and expenses of securities research) is essentially subsidized by the investment banking division of the integrated broker-dealer firm. The problematic result is at the same time to subsidize and arguably distort securities research.
This point distinguishes the securities analysts from the accountant. That is, if the auditor is prohibited from consulting for the client, both the auditing and the consulting function will survive. But, in particular because the costs of securities research cannot be easily passed on to the retail customer, a Glass-Steagall divorce might imply that the number of securities analysts wold shrink by a substantial fraction.(14) A cynic might respond: why seek to maximize biased research? Yet, if the number of analysts were to fall by, hypothetically, one half, market efficiency might well suffer, and many smaller firms simply would not be regularly covered by any analyst. Hence, the divestiture approach may entail costs and risks that cannot be reliably estimated.
2. Piecemeal Reform: Policing Conflicts. Proposed Rule 2711 represents an approach of trying to police conflicts and prevent egregious abuse. The practical ability of regulators to do this effectively is always open to question. For example, although proposed Rule 2711 generally prohibits investment banking officials from reviewing research reports prior to publication, it does permit a limited review "to verify the factual accuracy of information in the research report" (see Rule 2711(b)(3)). It is easy to imagine veiled or stylized communications that signal that the investment banking division is displeased and will reduce the analyst's compensation at the next regular salary review. Such signals, even if they consists only of arched eyebrows, are effectively impossible to prohibit. Still, at the margin, intelligent regulation may curtail the more obvious forms of abuse.
Although proposed Rule 2711 addresses many topics, it does not address every topic. Some other topics that may merit attention are discussed below, but they are discussed in the context of suggesting that Congress might give the NASD general policy instructions and ask it to fine tune more specific rules that address these goals:
1. An Anti-Retaliation Rule. According to one survey,(15) 61% of all analysts have experienced retaliation - - threats of dismissal, salary reduction, etc - - as the result of negative research reports. Clearly, negative research reports (and ratings reductions) are hazardous to an analyst's career. Congress could either adopt, or instruct the NASD to adopt, an anti-retaliation rule: no analyst should be fired, demoted or economically penalized for issuing a negative report, downgrading a rating, or reducing an earnings, price, or similar target. Of course, this rule would not bar staff reductions or reduced bonuses based on economic downturns or individualized performance assessments. Thus, given the obvious possibility that the firm could reduce an analyst's compensation in retaliation for a negative report, but describe its action as based on an adverse performance review of the individual, how can this rule be made enforceable? The best answer may be NASD arbitration. That is, an employee who felt that he or she had been wrongfully terminated or that his or her salary had been reduced in retaliation for a negative research report could use the already existing system of NASD employee arbitration to attempt to reverse the decision. Congress could also establish the burden of proof in such litigation and place it on the firm, rather than the employee/analyst. Further, Congress could entitle the employee to some form of treble damages or other punitive award to make this form of litigation viable. Finally, Congress could mandate an NASD penalty if retaliation were found, either by an NASD arbitration panel or in an NASD disciplinary proceeding.
2. A No-Selling Rule. If we wish the analyst to be a more neutral and objective umpire, one logical step might be to preclude the analyst from direct involvement in selling activities. For example, it is today standard for the "star" analyst to participate in "road shows" managed by the lead underwriters, presenting its highly favorable evaluation of the issuer and even meeting on a one-to-one basis with important institutional investors. Such sales activity seems inconsistent with the much-cited "Chinese Wall" between investment banking and investment research.
Yet, from the investment banking side's perspective, such participation in sales activity in what makes the analyst most valuable to the investment banker and what justifies multimillion dollar salaries to analysts. Restrict such activities, they would argue, and compensation to analysts may decline. Of course, a decline in salaries for super-stars does not imply a reduction in overall coverage or greater market inefficiency.
Although a "no-selling" rule would do much to restore the objectivity of the analyst's role, one counter-consideration is that the audience at the road show is today limited to institutions and high net worth individuals. Hence, there is less danger that the analyst will overreach unsophisticated retail investors. For all these reasons, this is an area where a more nuanced rule could be drafted by the NASD at the direction of Congress that would be preferable to a legislative command.
3. Prohibiting the "Booster Shot." Firms contemplating an IPO increasingly seek to hire as lead underwriter the firm that employs the star analyst in their field. The issuer's motivation is fueled in large part by the fact that the issuer's management almost invariably is restricted from selling its own stock (by contractual agreement with the underwriters) until the expiration of a lock-up period that typically extends six months from the date of the offering. The purpose of the lock-up agreement is to assure investors that management and the controlling shareholders are not "bailing out" of the firm by means of the IPO. But as a result, the critical date (and market price) for the firm's insiders is not the date of the IPO (or the market value at the conclusion of the IPO), but rather the expiration date of the lock-up agreement six months later (and the market value of the stock on that date). From the perspective of the issuer's management, the role of the analyst is to "maintain a buzz" about the stock and create a price momentum that peaks just before the lock-up's expiration.(16) To do this, the analyst may issue a favorable research report just before the lock-up's expiration (a so-called "booster shot" in the vernacular). To the extent that favorable ratings issued at this point seem particularly conflicted and suspect, an NASD rule might forbid analysts associated with underwriters from issuing research reports for a reasonable period (say, thirty days) both before and after the lock-up expiration date. Proposed Rule 2711 stops well short of this and only extends the "quiet period" so that it now would preclude research reports for this first 40 days after an IPO. Such a limited rule in no way interferes with the dubious tactic of "booster shots."
4. Summary: The most logical and less overbroad route for Congress to take with regard to securities analysts and their conflicts is to pass legislation giving the NASD more specific guidance and instructions about the goals that they should pursue and then instruct the NASD to conduct the necessary rule-making in order to fine tune this approach. NASD penalties might also properly be raised. This approach spares Congress from having to adopt a detailed code of procedure, avoids inflexibility and rigid legislative rules, and relies on the expertise of the SEC and the NASD, as paradigms of sophisticated administrative agencies.
1. I summarize much of this literature and the absence of any meaningful effort at internal self-discipline in a recent article. See Coffee, The Acquiescent Gatekeeper: Reputational Intermediaries, Auditor Independence and the Governance of Accounting (2001). This article, written well before the Enron story broke, is available on the Social Science Research Network ("SSRN") at www.ssrn.com at id=270994.
2. See George Moriarty and Philip Livingston, "Quantitative Measures of the Quality of Financial Reporting," 17 Financial Executive 55 (July 1, 2001).
3. Id. 715 of these restatements involved Nasdaq listed companies; 228 involved New York Stock Exchange companies; the rest were listed either on the American Stock Exchange or traded in the over-the-counter market. Premature revenue recognition was found to be the leading cause of restatements.
4. 17 CFR § 201.102. The SEC's authority under Rule 102(e) was clouded by the D.C. Circuit's decision in Checkosky v. SEC, 139 F.3d 221 (D.C. Cir. 1998) (dismissing Rule 102(e) proceeding against two accountants of a "Big Five" firm). The SEC revised Rule 102 in late 1998 in response to this decision (see Securities Act. Rel. No. 7593 (Oct. 18, 1998), but its authority in this area is still subject to some doubt that Congress may wish to remove or clarify.
5. A December, 2000 Thomson Financial Survey reported that 71% of all analyst recommendations were "buys" and only 2.1% were sells. Apparently, only 1% of 28,000 recommendations issued by analysts during late 1999 and most of 2000 were "sells." This study also finds that the overall "buy" to "sell" ratio shifted from 6:1 in the early 1990's to 100:1 by sometime in 2000. Of course, this shift also coincided with the Nasdaq bull market of the 1990's.
6. See R. Michaely and K. Womack, Conflict of Interest and the Credibility of Underwriter Analyst Recommendations, 12 Review of Financial Studies 653 (1999).
7. See S. Barr, "What Chinese Wall," CFO Magazine, March 1, 2000.
8. H. Lin and M. McNichols, Underwriting Relationships and Analysts' Earnings Forecasts and Investment Recommendations, 25 J. of Accounting and Economics 101 (1997).
9. J. Cote, Analyst Credibility: The Investor's Perspective, 12 J. of Managerial Issues 351 (Fall 2000).
10. D. Galant, "The Hazards of Negative Research Reports," Institutional Investor, July 1990.
11. Laura Unger, "How Can Analysts Maintain Their Independence?" Speech at Northwestern Law School (April 19, 2001).
12. See File No. SR-NASD-2002-21 (February 8, 2002).
13. See The Glass-Steagall Act of 1933, 12 U.S.C. § 36 et. seq. (separating commercial and investment banking).
14. I recognize that the number of "buy side" analysts employed by institutional investors might correspondingly increase, but not, I think, to a fully compensating degree. Moreover, "buy side" analysts do not publish their research, thus implying increased informational asymmetrics in the market.
15. See Galant, supra note 10, and Cote, supra note 9.
16. This description of the analyst's role (and of the underwriter's interest in
attracting "star" analysts) essentially summarizes the description given by three
professors of financial economics, Rajesh Aggarwal, Laurie Krigman and Kent
Womack, in their recent paper, Aggarwal, Krigman, and Womack, Strategic IPO
Underpricing, Information Momentum, and Lockup Expiration Selling (April
2001) (available on SSRN).
Home | Menu | Links | Info | Chairman's Page