read article and answer question

E V A L U A T I N G S O X

A considerable body of corporate finance and accounting research analyzes the efficacy of the substantive corporate governance mandates of sox. The data do not support the view that the legislation will improve corporate governance or performance.

I N D E P E N D E N T AU D I T C O M M I T T E E S Section 301 of sox requires all listed companies to have audit committees com- posed entirely of independent directors, as defined by Con- gress. The rationale for the rule is that such directors can be expected to be effective monitors of management, and thereby reduce the possibility of audit failure, because their financial dependence on the firm is limited to directors’ fees. (Misstating earnings will not, for example, increase inde- pendent directors’ income as could be the case for insiders with bonus compensation related to earnings.) Congress also mandated disclosure of whether any of those directors are financial experts, along with an explanation—for firms with no expert on the audit committee—of why no com- mittee members are experts.

A large literature has developed on whether independent boards of directors improve corporate performance. Across a variety of analytical approaches, independent boards do not improve performance. Boards with too many outsiders may, in fact, have a negative impact on performance. There are fewer studies—four as of this writing—of the relation between audit committee composition and firm performance. None have found any relation between audit committee independence and performance, despite using a variety of performance measures, including accounting and market measures as well as measures of investment strategies and productivity of long-term assets.

Twelve studies have examined the impact of the independ- ence of audit committees on the probability of financial state-

36 R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6

C O R P O R A T E G O V E R N A N C E

Roberta Romano is the Oscar M. Ruebhausen Professor of Law at Yale Law School,

where she specializes in financial market regulation and corporate law and finance. She

may be contacted by e-mail at [email protected].

This article is condensed from Romano’s “The Sarbanes-Oxley Act and the Making of

Quack Corporate Governance,” Yale Law Journal, 2005.

Sarbanes-Oxley may have satisfied a political need, but it will do little to protect investors or strengthen the market.

Quack Corporate Governance

B Y R O B E RTA R O M A N O Yale Law School

he sarbanes- oxley act of 2002 (sox), in which Congress introduced a series of corporate governance initiatives into the federal securities laws, is not just a considerable change in law, but also a departure in the mode of regulation. The federal regime until then had consisted

primarily of disclosure requirements rather than substan- tive corporate governance mandates. Those mandates were left to the state corporate law. Federal courts had enforced such a view of the regime’s strictures by characterizing efforts of the Securities and Exchange Commission to extend its domain into substantive corporate governance as beyond its jurisdiction. sox alters that division of authori- ty by providing explicit legislative directives for sec regu- lation of what was previously perceived as the states’ exclu- sive jurisdiction.

sox was enacted in a flurry of congressional activity in the run-up to the midterm 2002 congressional elections after the spectacular failures of the once highly regarded firms Enron and WorldCom. Those firms entered bankruptcy proceedings in the wake of revelations of fraudulent accounting practices and executives’ self-dealing transactions. But many of the sub- stantive corporate governance provisions in sox are not reg- ulatory innovations devised by Congress to cope with defi- ciencies in the business environment in which Enron and WorldCom failed. Rather, they may more accurately be char- acterized as recycled ideas advocated for quite some time by corporate governance entrepreneurs.

T

Romano.final 12/8/05 5:40 PM Page 36

R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6 37

ment misconduct rather than performance. Of the 16 total studies of audit committee independence, 10 (including the four studies of explicit performance measures already noted) find that complete independence of the audit committee does not improve performance—whether performance is measured conventionally or by the existence of accounting impropri- eties—and one study reports inconsistent results (under one model formulation, independence improves performance, but not under all other models tested).

The data are mixed on whether even a committee with a majority of independent directors improves performance. But the issue for sox is whether complete independence improves on the effect of a majority-independent committee, not the effi- cacy of a majority of independent directors. A few studies find that having a director with financial expertise improves per- formance and, more specifically, that complete independence is less significant than expertise with respect to the relation between audit committee composition and accounting state- ment quality. Those results are notable in that sox does not mandate the presence of a financial expert on the audit com- mittee (it has only a disclosure requirement regarding financial expertise on the committee), while it does mandate completely independent audit committees.

Even when studies demonstrate positive statistical rela- tionships, they may be the result of selection effects rather than causality. For example, the finding of statistical signif- icance for director expertise in relation to financial statement restatements can be considered evidence that directors with expertise are effective monitors of accounting controls and audit quality—a rationale for reforming corporate gover- nance. But it is also possible that firms that are better man- aged, and hence less likely to restate their financial state- ments, choose to have independent directors with expertise. That is, a finding of significance may be a function of self- selection and not of the efficacy of the corporate governance mechanism. Accordingly, if selection effects explain the few studies with significant results, then that would strengthen the case against the mandate.

The literature on the composition of audit committees, in short, does not support the proposition that requiring audit committees to consist solely of independent directors will reduce the probability of financial statement wrongdoing or otherwise improve corporate performance. More important- ly, such results are found in the studies using the more sophis- ticated techniques. And even with properly specified statisti- cal tests, false positives—statistically significant results—can be expected five percent of the time, even though there is no significant relation between variables. The significant results in a small number of papers could well be false positives, the product of random error.

N O N AU D I T S E RV IC E S Section 201 of sox prohibits account- ing firms from providing specified nonaudit services to firms that they audit. The banned services include financial infor- mation system design and implementation, appraisal or valu- ation services, internal auditing services, investment banking services, legal and expert services unrelated to the audit, bro-

kerage services, and actuarial services. The rationale for the ban was that the receipt of high fees for nonaudit services com- promises auditor independence by providing auditors with a financial incentive to permit managers to engage in question- able transactions or accounting practices in the audit.

Because the sec had wanted to eliminate the provision of nonaudit services by auditors for some time prior to sox, numerous studies sought to ascertain whether the provision of such services by the external auditor compromises audit quality (the rationale advanced for banning the practice). Nine- teen of 25 studies suggest that sox’s prohibition of the purchase of nonaudit services from an auditor is unnecessary.

The majority (15) find no connection between the provision of nonaudit services and audit quality. One finds no connec- tion when the auditors are the Big Five accounting firms (including Arthur Andersen). And three find that nonaudit services improve audit quality (and two of the 15 that find no relation also find that audit quality improves in at least one model specification), which directly contradicts the rationale for the sox prohibition.

Of the remaining six studies, five find that audit quality is compromised, while one finds that audit quality is compro- mised in only one of several model specifications. However, the results of the initial and leading study by Richard Frankel, Mar- ilyn Johnson, and Karen Nelson, which found that audit qual- ity (measured by abnormal accruals) is compromised by the purchase of nonaudit services, are not robust. Numerous stud- ies have redone the analysis, refining the model in a variety of ways such as controlling for factors known to affect the audit performance measure used in the original study and using auditor independence measures that take account of the impor- tance of the client to the auditor. When the model is refined by any of those methods, the original results are not replicated. As a consequence, valid policy inferences cannot be drawn from Frankel, Johnson, and Nelson’s study. This could also be true for the other studies finding a significant inverse relation between nonaudit fees and audit quality. Less prominent than the Frankel et al. study but using the same methodology, those studies have not been the objects of further research.

The conclusion that audit quality—and hence auditor inde- pendence—is not jeopardized by the provision of nonaudit services is compelling not only because it is the finding of the vast majority of studies but also because it is the result of the studies using the most sophisticated techniques, as well as those whose findings are most robust to alternative model specifications. The absence of a systematic inverse relation between nonaudit fees and audit quality (across all measures of audit quality) in the scholarly literature is consistent with the Panel on Audit Effectiveness’s failure to identify a single instance of a compromised audit by auditors providing nonau- dit services in its field study of auditor independence. That find- ing no doubt contributed to the panel’s decision, as well as to that of the Independence Standards Board, not to recommend banning the provision of nonaudit services and to opt instead for bolstering the audit committee function by proposing that audit committees be composed of independent and financial- ly literate directors.

Romano.final 12/8/05 5:40 PM Page 37

C O R P O R A T E G O V E R N A N C E

EX EC U T I V E LOA N S Section 402(a) of sox prohibits corpora- tions from arranging or extending credit to executive officers or directors unless the corporation is a financial institution offering credit in the ordinary course of business and the terms of the credit are the same as those offered to the public. Loans became a focus of congressional attention in the wake of dis- closures that executives at Enron, WorldCom, Tyco Interna- tional, and Adelphia Communications had obtained extreme- ly large loans (in some cases in the hundreds of millions of dollars), personally benefiting from firms whose shareholders and employees suffered devastating financial losses. The blan- ket prohibition has engendered concern among practitioners

because it appears to prohibit standard compensation practices thought to be uncontroversial and beneficial.

In contrast to other sox corporate governance provisions, this initiative had not been a component of recent policy dis- cussions. The permissibility of such transactions had been set- tled state law for decades without generating scholarly con- troversy. It is not surprising that there is an absence of empirical research on the practice.

Motivated by the spotlight thrown on executive loans in the scandals leading to sox and by its ban on the practice, a recent study sought to measure the efficacy of executive loans by ana- lyzing whether they accomplish the purpose of increasing managerial stock ownership, thereby aligning managerial incentives with shareholder interests. Most sample loans were made to assist in stock purchases and stock option exercises, with a much smaller set consisting of relocation loans. The data are consistent with the hypothesis that most loans’ purpose is one of incentive alignment: There is an increase in executives’ equity ownership after the extension of credit to purchase stock

or to exercise stock options, although the increase is small rel- ative to loan value. Because executive loans increase manage- rial stock ownership, thus aligning managers’ and sharehold- ers’ interests, the blanket prohibition of executive loans in sox is self-evidently a public policy error. The provision in the orig- inal Senate bill (as well as the House bill), which was consistent with the conventional federal regulatory approach, required disclosure of executive loans but did not prohibit them.

EXECUTIVE CERTIFICATION Section 302 of sox requires the ceo and cfo to certify that the company’s periodic reports do not contain material misstatements or omissions and

“fairly present” the firm’s financial condition and the results of operations. The certification requirement contains sub- stantive corporate governance mandates. It imposes on the signing officers the responsibility for establishing and main- taining internal controls and for evaluating the effectiveness of those controls, along with the duty to disclose to the audit committee any deficiencies in the internal control design or any fraud involving any officer or employee with a significant role in the company’s internal controls. The officers’ signa- ture certifies both the undertaking of those tasks and the veracity of the financial information.

Prior to the enactment of sox, the sec imposed a certi- fication requirement on the largest public firms. This requirement was one of the proposals advanced by President Bush in his response to the Enron scandal, a 10-point plan to make corporate executives more accountable to investors. But even before the promulgation of the sec rule, ceos and cfos had always been required to sign the annu- al report and were liable for knowingly filing fraudulent

38 R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6

Romano.final 12/8/05 5:40 PM Page 38

reports as well as for inadequate internal controls. Two studies have sought to measure the efficacy of the sec’s

rule requiring executive certification of the financials of the largest firms by examining stock price reactions to timely and untimely certifications. The research question is whether the sec certification requirement provided new information to investors about firms’ financial conditions—as the literature puts it, was the requirement “value relevant”?—and more specif- ically, did a failure to comply, or early compliance, provide infor- mation to investors? One study finds that the certification requirement had no impact. But the other study finds that for a subset of firms considered to be informationally opaque (bank

holding companies), early certification provided new, and pos- itive, information to the market.

The contrary findings of the two event studies of the certi- fication requirement do not allow any definitive conclusion regarding the efficacy of the provision for improving the abili- ty of investors to distinguish between high- and low-quality firms. One policy approach that would reconcile the results would be to render the certification regime optional. That would permit firms for which there is a benefit to engage in special cer- tifications rather than requiring the conventional financial state- ment signatures (for example, opaque firms such as bank hold- ing companies) to do so. Firms would select into the regime when the burden of compliance was more likely to produce a positive payoff to their investors.

T H E P O L I T I C A L E C O N O M Y O F S O X

The empirical literature suggests that the principal corporate governance mandates in sox will not benefit investors. Why did Congress enact mandates that will not improve investor

welfare? Although much of the research reviewed in this arti- cle was not available to Congress during its deliberations, the findings on independent audit committees and nonaudit serv- ices were. But the literature was not even cursorily addressed during the legislative process.

The corporate governance mandates stemmed from the interaction of the Senate Banking Committee chairman’s response to the suggestions of policy entrepreneurs and party politics in an election cycle coinciding with spectacu- lar corporate scandals, a sharp stock market decline, and the consequent political collapse of the interest groups (the accounting profession and the business community) whose

policy position was most consistent with the empirical lit- erature. sox exemplifies low-quality legislative decision- making in the context of a crisis, a feature that has been repeated on other occasions when the federal government has intervened in financial markets.

Even with a committee system permitting specialization, legislators cannot be expected to have extensive technical expertise. Therefore, some of the shortcomings of sox’s cor- porate governance mandates should be assigned to legislative staff. Whether that failure was the result of staff members’ ide- ological commitments, a lack of the technical skill necessary to evaluate the literature, or a combination of the two is unknown. But members of Congress select their staff, and in that regard they bear responsibility for the poor performance of those individuals.

B AC KG R O U N D sox was adopted in July 2002, slightly less than a year after the Enron scandal broke. The Enron scan- dal was followed by revelations of accounting fraud and

R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6 39

M O

R G

A N

B A

L L

A R

D

Romano.final 12/8/05 5:40 PM Page 39

C O R P O R A T E G O V E R N A N C E

insider self-dealing at several large corporations, nearly all of which were thereafter pushed into bankruptcy: Adelphia Communications, Global Crossing, Tyco International, and WorldCom. The stock market declined sharply throughout the time frame in which Congress was considering the sox legislation. The low point, which represented more than a one-third loss in value of the index over the preceding year, occurred on July 23rd, the day before the conference com- mittee reported out a bill and the second trading day after the bankruptcy filing of WorldCom.

Members of Congress, not surprisingly, were attentive. Senators explicitly referred to the steep stock market decline in July as a rationale for the need for legislative action. That

response was certainly not out of the ordinary. As Stuart Banner notes, most new major securities regulation in the United States, as well as the United Kingdom, has followed stock market crashes. Congress, with an election looming, interpreted the market decline from April through July 2002 as requiring legislative action.

The corporate governance mandates were neither a prin- cipal nor a subsidiary focus of legislative consideration. With the exception of the restriction on the provision of nonau- dit services by auditors, for all practical purposes they were not even discussed.

The usually key role of committees in the formulation of leg- islation was virtually absolute in this case, and the Democrats’ drafting in the Senate Committee was heavily informed by the views of former sec chairman Arthur Levitt and his former chief accountant, Lynn Turner. In a remarkable turn of events, Levitt was able to revive his agenda for accounting regulation (particularly the prohibition on nonaudit services), which had failed less than two years earlier when confronted with bipar- tisan congressional support for the accounting profession’s position against Levitt’s proposals.

H O U S E D E B AT E The majority party exercises strict control over the legislative process in the House, and the adoption of the bill of the Financial Services Committee chaired by Rep. Michael Oxley (R, Ohio) was no exception: The Repub- lican Party shepherded the bill through the floor with one day of debate.

No one in the House debate on Rep. Oxley’s bill men- tioned audit committee independence or executive loans, the subjects of the sox corporate governance mandates most intrusive on state law jurisdiction, nor did those mandates appear in House Democrats’ bills. In addition, none of the

governance provisions that had been introduced in the Sen- ate bill were even mentioned in the subsequent House debate over the Senate bill.

Political scientists have characterized House floor debate as for “public consumption” rather than for persuasion of members on the other side of an issue. Even from that per- spective, the lack of reference to the corporate governance reforms that were included in the final bill is notable because it indicates that members of Congress did not consider those provisions to be matters that would serve either to justify their votes or to demonstrate to constituents how legislation was solving the “Enron problem.” The governance provi- sions therefore would appear to have been of principal inter-

est to corporate governance policy entrepreneurs—indi- viduals “inside the Beltway”—at least as far as House members were concerned.

S E N AT E D E B AT E Initially, Enron’s collapse in the fall of 2001 generated a crisis situation and a media frenzy, as every con- gressional committee that could find some jurisdictional basis held a hearing on the scandal. But by April, the sense of an emergency had lessened such that the members of the Senate Banking Committee did not feel any urgency to agree on a bill in response to the House action. Indeed, even after Sen. Paul Sarbanes (D, Md.) took several months to craft a bill that met bipartisan committee approval, it appeared that the bill would not progress. The best that Senate Majority Leader Thomas Daschle (D., S.D.) could do was to try to schedule a vote on the bill for sometime after the August recess, and leg- islators opposed to the bill expressed the view that “Enron’s moment as a galvanizing issue has quickly passed.”

When the WorldCom scandal broke on June 26, the polit- ical environment changed dramatically once again and Daschle, now predicting 80 votes in support of the bill, was able to move it up on the calendar for a July vote. Senator Phil Gramm (R, Texas), the ranking Republican on the commit- tee who opposed the bill and had earlier thought the feeding frenzy was over and the movement for legislation stopped, now did not even attempt to stem the bill’s progress to the floor and a vote. The Senate moved on the legislation rapid- ly, agreeing to cloture after having taken no action on the House bill for months.

Only one of the corporate governance mandates (the restriction on nonaudit services) was the focus of significant debate on the Senate floor. The debate over the nonaudit services prohibition was, in large part, a replay of a battle

40 R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6

SOX exemplifies low-quality legislative decision-making in the context of a crisis, a feature

that has been repeated on other occasions.

Romano.final 12/8/05 5:40 PM Page 40

R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6 41

over the regulation of the accounting industry fought two years earlier when Levitt was sec chair. But the environment this time was markedly different. There was a media frenzy, heightened by a sharply declining stock market and high- profile accounting frauds and business failures, in the mid- dle of an election year. In this charged atmosphere, Levitt’s earlier reform proposals now seemed prescient (at least to the Democrats for whom Levitt was a source of expertise), and the accounting industry had lost its public credibility with the audit failures.

There was a near-total absence of discussion on the Sen- ate floor of the two corporate governance mandates—the independent audit committee provisions and the certification requirement—that were included in the Senate but not the House bill. Legislators perceived those provisions as unprob- lematic. Only 26 senators referred to any of those provisions on the floor, and nearly all of those references were part of laundry lists in which senators expressed support for the leg- islation by enumerating specific provisions in the bill.

In the limited time for consideration of the bill following the cloture motion, a few amendments agreed to by both par- ties were added on the floor without debate, including the third governance mandate absent from the House bill, pro- hibition of loans to executives. There was no discussion of that amendment when it was offered by Sen. Charles Schumer (D, N.Y.): It was immediately unanimously agreed to without a roll call vote. A few days earlier in a speech on Wall Street, President Bush had called on corporate boards to prevent officers from receiving company loans. Schumer referred to the president’s remarks when introducing the amendment and noted that he had “spoken to the people in the White House who were supportive of [the] amendment.”

Just why the president made the suggestion is unknown. Perhaps he was seeking to immunize himself from criticism of loans that he had received when he was in business. But whatever the reason, his remarks appear to have been a deci- sive factor in the inclusion of this provision because such a provision had previously been rejected by the Banking Com- mittee. The near-total absence of considered discourse on sox’s governance provisions in the Senate is consistent with the characterization of the corporate governance issues as being “below the radar screen” and “inside the Beltway.” In the limited time frame available for legislative debate, sen- ators did not focus any attention on the corporate gover- nance provisions.

Thus, as in the House, legislators who could not possibly be informed on technical issues and who felt that they had to act under the pressure of mounting corporate account- ing scandals simply accepted the bill that was presented. That bill consisted of measures advocated by policy entre- preneurs (former government officials aligned largely with one political party) as filtered by the Banking Committee chairman. Many of those individuals were advancing pro- posals that they had previously advocated and that they believed would improve the quality of financial reporting, despite a virtually complete lack of data supporting their beliefs. With little attention accorded to the proposals in the

committee hearings and even less attention on the floor, the disjuncture between the recommended policies and the empirical literature was never even acknowledged, as might have been possible if the legislative process had not been operating in a crisis atmosphere.

The policy entrepreneurs on whom the Democrats relied in the context of the highly publicized and time-restricted delib- eration over sox—former sec chairman Levitt and former chief accountant Turner—are the key to understanding why Congress enacted a series of provisions that are ill-matched to fulfill their stated objectives. During Levitt’s term as chairman, empirical research was accorded little weight in the setting of regulation. This fact is made plain by the Levitt sec’s response to the Panel on Audit Effectiveness’s failure to find that the pro- vision of nonaudit services compromised audit quality. In the release on the proposed auditor independence rules restrict- ing nonaudit services, the agency summarily dismissed the concern raised by the accounting profession that, in light of the panel’s report, there was no evidence of a connection between the provision of nonaudit services and accounting fraud or audit compromise.

The sec stated, “Studies cannot always confirm what common sense makes clear.” The panel, it should be recalled, was created at Levitt’s request. Not surprisingly, a statute informed by Levitt’s perspective would not be responsive to the concerns of a literature that did not fit with his preconceptions.

C O N F E R E N C E C O M M I T T E E Three factors—a media frenzy, the precipitous drop in the stock market, and reelection con- cerns—led the conference committee to act quickly and report a bill virtually identical to the Senate bill, with only a few minor changes (such as inclusion of the House’s length- ier criminal sanctions). That is, the Republicans capitulated to the Democrats’ bill. As House Minority Leader Richard Gephardt (D, Mo.) put it, the Republicans’ action was “an unconditional surrender.” This may well have been a prudent decision for Republicans from the perspective of their elec- toral ambitions. As commentators have suggested, the elec- toral gains Republicans made in the 2002 election were the result of the public’s concern about national security rather than corporate scandals, which were thought to be an issue favoring the Democrats. The enactment of sox may have contributed to a shift in public focus by removing corporate scandals from the public policy agenda.

T H E R O L E O F P O L I C Y E N T R E P R E N E U R S

Given the general lack of interest in the sox corporate gover- nance mandates shown by legislators during the floor debate, public policy entrepreneurs (who were mostly former gov- ernment officials) and Senate Banking Committee chairman Sarbanes were key formulators of sox’s corporate governance provisions. Virtually all of those individuals were associated in some capacity with the sec.

EX EC U T I V E LOA N S The origin of the executive loan provi- sion in the Senate bill is the easiest of the corporate gover-

Romano.final 12/8/05 5:40 PM Page 41

C O R P O R A T E G O V E R N A N C E

nance mandates to trace. At the initial Senate hearing, one witness expressed concern about executive loans. That was former sec chairman Richard Breeden, who recommended that all loans be disclosed in corporate proxies and, when above a specified amount, subject to shareholder approval. This resonated with Sen. Sarbanes, who proceeded to ask six other witnesses (witnesses on two panels considered to have expertise in corporate governance) what they thought of Breeden’s testimony regarding loans. Only one witness, for- mer Ohio senator Howard Metzenbaum representing the Consumer Federation of America, thought that loans to offi- cers should be banned. The other witnesses queried—a prominent corporate governance attorney and representa- tives of institutional investors and the AFL-CIO—expressed support solely for a disclosure provision. Indeed, one of the witnesses noted that company loans originated for the legit- imate purpose of assisting relocations and argued that it would get “very messy” if Congress were to say that “you can- not ever lend money to an employee.”

Breeden was also a witness at the House hearing on the minority bill, but he did not mention the issue of executive loans in his House testimony. Because his testimony to the House occurred two months after he had testified to the Sen- ate, whatever the reason for the omission, it was not because the issue had not occurred to him. The disclosure provision was amended on the floor to become a prohibition, in the wake of the president’s remarks. It is ironic that the avenue facilitating its inclusion—the loan disclosure provision—was an idea that appealed more to the committee chairman than to its origi- nator, Richard Breeden, for whom it was one, and in all likeli- hood not the most important, of a series of proposals, most of which were not pursued by the committee.

I N D E P E N D E N T AU D I T C O M M I T T E E S The origin of the Sen- ate provision requiring independent audit committees is a bit harder to trace than that of the loan provision. The composi- tion of the audit committee was a concern emphasized by for- mer sec chairman Roderick M. Hills in both chambers’ earli- est hearings, although his specific proposal was to require that members of the audit committee be appointed by nominating committees consisting exclusively of independent directors. The initial stock exchange requirement of an audit committee occurred on his watch as sec chairman, in 1974, in the wake of a corporate scandal involving sensitive payments to foreign officials. Hills perceived his recommendation as being a time- ly and necessary follow-up to that legislation.

Other witnesses on the Senate panel with Hills also referred to the importance of independent audit committees or to a vague need to “enhance” their independence, but they did not provide specific proposals. In later sessions, however, witnesses made more concrete recommendations on independence sim- ilar to the provisions included in the Senate bill. Most notably, Turner stated that the stock exchange rules permitting excep- tions to the requirement that all audit committee members be independent should be eliminated. Another former sec chief accountant, Michael Sutton, also recommended requiring completely independent audit committees.

The third former sec chief accountant who testified on the panel, Walter Schuetze, stated that Enron’s problems were inherent to current accounting rules (that assets and liabilities are not marked to market) rather than lack of auditor inde- pendence or oversight. He also provided copies of his articles discussing how accounting ought to be reformed, one of which referred to another article’s “excellent discussion and analysis” of why the presence of independent audit committees cannot improve the quality of an audit. He did not, however, challenge his fellow panelists’ recommendations on audit committee composition, nor was he asked for his views on that matter, and the suggestion in his articles that independent audit commit- tees would not alleviate the problem was not picked up by any senator. It was simply ignored.

The audit committee independence idea had been advanced by former high-ranking government officials who were well regarded by many members of the Senate Banking Committee. The committee chairman found the idea attractive, and the committee never had to confront the inconvenient reality that there was a relevant literature whose learning was starkly at odds with this regulatory focus. As far as the committee was concerned, the literature did not exist.

Again, a comparison with the more focused House hear- ings is instructive. In the House hearings, only a few wit- nesses raised the issue of audit committee independence, and none advocated requiring a majority of independent direc- tors on the board. No doubt, the difference in testimony and emphasis on audit committee independence across the chambers reflects the difference in party control. This was not a top concern of Republicans in the House, and the wit- nesses they called either were also not interested in the issue or determined it was best to direct their attention to matters the majority deemed a priority. In fact, even the ranking Democrat, Rep. John LaFalce (N.Y.), who considered reform of boards’ and audit committees’ independence a top prior- ity, indicated that he believed legislation unnecessary because committee independence was within the sec’s rule- making authority. Accordingly, the difference in agenda con- trol and dynamics across the chambers on the issue of audit committee independence sheds light on the difference in the content of the chambers’ bills: No witnesses before the House explicitly advocated legislation on independent audit com- mittees, fewer witnesses raised the issue there than in the Senate, and the House committee chairman did not latch onto the idea as worthy of pursuit.

EX EC U T I V E C E RT I F ICAT IO N In the Senate, former sec chief accountant Turner was the first to recommend mandatory executive certification of financial statements, which he noted was a practice followed in foreign jurisdictions. Thereafter, three other witnesses expressed support for a certification requirement as an incentive device to improve reporting. Those endorsements were volunteered because Sen. Sarbanes did not seek other witnesses’ views on Turner’s proposal. Sarbanes’s lack of follow-up on Turner’s suggestion may well have been a function of a lack of interest in the recommendation. The cer- tification requirement was, in fact, the one governance man-

42 R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6

Romano.final 12/8/05 5:40 PM Page 42

date to which Sarbanes did not refer in his remarks on the Sen- ate floor during the deliberations on sox.

A week after Turner’s testimony, President Bush announced a 10-point plan for improving corporate respon- sibility, which included a similar certification requirement. The sec also indicated that it intended to implement the pro- posal on its own. Those comments were critical to the certi- fication requirement’s inclusion in the committee bill, given Sarbanes’s low level of personal interest in it. The legislative history notes that the bill “in effect” adopted Bush’s propos- al, while crediting the precise formulation to Sen. Zell Miller

(D, Ga.), who was a crucial committee vote in Sarbanes’s effort to produce a bipartisan bill.

In contrast to the Senate, only one witness at the House hearings raised the issue of executive certification. That wit- ness was once again Turner, who now endorsed the Bush administration’s suggestion of certification in response to questions by ranking member LaFalce on how to improve auditor independence and on the need to restructure audit committees. The House hearing was held after the president had announced his corporate responsibility proposals, but also after the Republicans had drafted their bill, which did not include a certification provision. Because the president’s proposal did not require legislative action—the sec could (and did) implement it under its own rulemaking authority— the House Republicans did not have to amend their bill for the proposal to move forward.

N O N AU D I T S E RV IC E S The policy entrepreneur supporting the nonaudit services provision was Arthur Levitt, who led the sec’s initiative on the issue two years before. Levitt was able to advance his agenda of a total ban on the provision of nonaudit services by auditors now that the accounting pro- fession had landed in Congress’s crosshairs. Levitt and Turn- er displayed the skills of expert agenda-setting entrepreneurs. Through their testimony during the hearings (and addition- al offstage communication, including considerable media exposure), they were able to link the scandal with Levitt’s position on auditors’ provision of consulting services and with the accounting profession’s successful opposition to his agenda to ban such services while he was sec chairman. Members of Congress who had supported the accounting industry against Levitt’s efforts to ban nonaudit services in the rulemaking process less than two years earlier hastily aban- doned that position in the aftermath of Enron.

Three of the witnesses who opposed expanding the restric-

tions on nonaudit services supported their position by not- ing that there was no evidence that the provision of nonau- dit services compromises audit quality. But the passing ref- erences were ignored.

The adoption of the nonaudit services restriction illus- trates the critical entrepreneurial role of the committee chair- man. With the bulk of his career in the public sector and a very liberal voting record, Sen. Sarbanes’s priors would make him favorably disposed to greater regulation of business, such as the use of mandates rather than disclosure as the cor- porate governance approach for sox, and to adoption of a

nonaudit services prohibition that was stricter than the House’s version. It is altogether understandable that the few references to data inconsistent with the recommendations to restrict nonaudit services by witnesses such as Levitt, who for the most part shared Sarbanes’s worldview, did not enter into the senator’s calculation and influence his adoption of their recommendations.

In sum, the most active and influential policy entrepreneurs in the sox legislative process were former sec officials. The adoption of the sox corporate governance mandates are the successful culmination of a multi-decade effort by the sec’s personnel to assert authority over public corporations in areas long considered the jurisdiction of the states.

S Y M B O L I C P O L I T I C S O R W I N D O W D R E S S I N G ?

The sox corporate governance mandates were not carefully considered by Congress. In particular, they were not evaluat- ed in light of the empirical literature questioning their benefits.

Some have argued that sox prevented the passage of even more egregious legislation. That may be true, but sox does have costs. The compliance costs to meet the certification requirement appear to be considerable, especially for small- er firms. For example, a recent survey of companies’ pro- jected expenditures to meet the sox internal controls pro- visions by the financial officers’ professional organization shows that companies with annual revenues over $5 billion projected external consulting, software, and additional audit fees of $2.9 million per company, compared to a pre-adop- tion projection of $222,200 by companies with annual rev- enues under $25 million.

Taking the revenue thresholds as a benchmark, smaller com- panies’ projected outlays as a proportion of revenue are an order of magnitude greater than larger companies’ (0.009 com- pared to 0.0006). Another survey, of firms going private, report- ed that the cost of being public more than doubled after sox,

R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6 43

SOX was an inefficient way to calm the media frenzy over corporate scandals, even if more costly

governance proposals could be imagined.

Romano.final 12/8/05 5:40 PM Page 43

C O R P O R A T E G O V E R N A N C E

rising on average from $900,000 to $1.95 million, with the increase attributed primarily to higher audit, insurance, and outside-director fees. Those data indicate that sox imposed a far more significant burden on small than on large firms.

The study also finds that smaller-sized firms’ expenditures on directors’ compensation appear to have massively increased. It reports two measures of expenditures for a small sample of firms stratified by size: The cash compensation that medium- sized firms paid to outside directors increased from $21,688 to $40,783 between 2001 and 2004 (the effective date for com- pliance with most sox rules), and small firms’ compensation to outside directors increased from $7.25 per $1,000 in net sales to $9.76 over the same period, compared to a trivial increase ($0.20) for large firms. Furthermore, sox appears to have affected the rate at which small firms stay public.

In addition to direct compliance costs, there are some costs that are difficult to quantify but that could prove to be sub- stantial, such as the contraction in financing opportunities for small and mid-sized businesses, as public firms are deterred from acquiring private and foreign firms or as those firms do not go public because of the sox mandates. To the extent that acquirers’ transaction risk has increased because of the certi- fication requirement, the efficiency of the market for corporate control could be affected—a potentially serious and unin- tended cost of the legislation.

Finally, there are potential long-run costs for U.S. stock exchanges, as foreign firms shift to the principal competitor venue—the London exchange—to avoid sox. The cost and dif- ficulty for foreign firms of complying with sox’s requirements may well be greater than for smaller U.S. firms, or at least much less worthwhile when balanced against the benefit obtained from a U.S. listing. This is not simply speculation because many foreign firms are contemplating delisting. U.S. investors, as well as exchanges, would be disadvantaged by such a trend because, while they will still be able to purchase such firms’ shares abroad, transaction costs will be higher. (Besides higher trad- ing fees, the transactions will not be in U.S. dollars.)

In sum, given the available information, it is not credible to characterize sox’s governance mandates as no- or low-cost window dressing. sox was an inefficient way to calm the media frenzy over corporate scandals, even if more costly governance proposals could be imagined.

POLICY IMPLICATIONS Analysis of the empirical literature and the political dynamics relating to the sox corporate governance mandates indicates that those provisions were poorly conceived because there was no basis to believe they would be efficacious. Hence, there is a disconnect between means and ends.

The straightforward policy implication is that the mandates should be rescinded. The easiest mechanism for operational- izing such a policy change would be to make the sox mandates optional, i.e., statutory default rules that firms could choose to adopt. An alternative and more far-reaching approach, which has the advantage of a greater likelihood of producing the default rules preferred by a majority of investors and issuers, would be to remove corporate governance provisions com- pletely from federal law and remit those matters to the states.

Finally, a more general implication concerns emergency leg- islation. It would be prudent for Congress, when legislating in crisis situations, to include statutory safeguards that would facilitate the correction of mismatched proposals by requiring, as in a sunset provision, revisiting the issue when more con- sidered deliberation would be possible.

C O N C L U S I O N

This article has examined the substantive corporate gover- nance mandates adopted by Congress in the wake of the Enron scandals. An extensive empirical literature suggests that those mandates were seriously misconceived because they are not likely to improve audit quality or otherwise enhance firm per- formance and thereby benefit investors as Congress intended. In the frantic political environment in which sox was enact- ed, legislators adopted proposals of policy entrepreneurs with neither careful consideration nor assimilation of the literature at odds with the policy prescriptions.

The specific policy implication drawn from this article’s analysis of the scholarly literature and political dynamics is that the mandates should be rescinded, either by trans- forming them into statutory defaults that apply to firms at their option or by removing them completely and redirect- ing jurisdictional authority to the states. The more general implication is the cautionary note that legislating in the immediate aftermath of a public scandal or crisis is a formula for poor public policymaking (at least in the context of finan- cial market regulation). The high salience of events forestalls a careful and balanced consideration of the issues, providing a window for action by the better-positioned, not the better- informed, policy entrepreneurs.

This is a particular concern because legislation drafted in a perceived state of emergency can be difficult to undo. It took more than 60 years to repeal the Glass-Steagall Act, the New Deal financial market regulation that is now widely rec- ognized as having greatly contributed to the banking deba- cle of the 1980s. The problem would be mitigated by rou- tinizing the inclusion in emergency legislation of a provision for revisiting the legislation to determine whether continu- ation is warranted at a later date when more deliberative reflection is possible.

Congressional repeal of sox’s corporate governance mandates is not on the near-term political horizon. Office- holders would not want to be perceived as revising rules that are supposed to diminish the likelihood of corporate accounting scandals. The alternative of treating sox as a set of default rules could be implemented by the sec under its general exemptive authority, but it is improbable that the agency will do so in a comprehensive way, in part because it is still stinging from being perceived as lagging behind state regulators in finding and prosecuting entire financial industry sectors for alleged misconduct. It is therefore important to work to educate the media, the public, politi- cal leaders, and agency personnel regarding the reality that Congress committed a public policy blunder in enacting sox’s corporate governance mandates and that there is a need to rectify the error.

44 R E G U L AT I O N W I N T E R 2 0 0 5 - 2 0 0 6

R

Romano.final 12/8/05 5:40 PM Page 44