university of michigan€¦ · university of michigan retaining mandatory securities disclosure:...

UNIVERSITY OF MICHIGAN

RETAINING MANDATORY SECURITIES

DISCLOSURE: WHY ISSUER CHOICE

IS NOT INVESTOR EMPOWERMENT

Merritt B. Fox

UNIVERSITY OF MICHIGAN

LAW SCHOOL

PAPER # 99-008

A revised version of this working paper is forthcoming in the Virginia Law Review, 1999.)

This paper can be downloaded without charge at:

The Social Science Research Network Electronic Paper Collection:http://papers.ssrn.com/paper.taf?abstract_id=155928

TABLE OF CONTENTS

I. THE NATURE OF THE DISCLOSURE REGIMES ISSUERS WILL CHOOSE . . . 5A. The Persons Choosing an Issuer’s Regime: Private versus

Social Optimality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6B. Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1. The Divergence of Private from Social Costs . . . . . . 82. Romano’s Dismissal of the Importance of the

Private/Social Cost Divergence . . . . . . . . . . . . . 93. The Further Divergence of Private from Social Cost

Because of Management’s Agency Relationshipwith the Shareholders . . . . . . . . . . . . . . . . . . . . 16

C. Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171. Reduction in investor risk . . . . . . . . . . . . . . . . . . . . 182. Improved project choice . . . . . . . . . . . . . . . . . . . . . 183. Greater Managerial Adherence to Shareholder

Interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

II. EMPIRICAL EVIDENCE CONCERNING THE WELFARE EFFECTS OF

MANDATORY DISCLOSURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28A. Empirical Evidence Concerning the Imposition of Mandatory

Disclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281. Studies of the Effect of New Issue and Periodic

Mandatory Disclosure on Price Accuracy . . . . 282. Studies of the Effect of New Issue Mandatory

Disclosure on Rate of Return. . . . . . . . . . . . . . . 383. Studies of the Effect of Periodic Mandatory Disclosure

on Rate of Return . . . . . . . . . . . . . . . . . . . . . . . 42B. The State Corporate Law Competition Literature . . . . . . . 47C. Implications of the Available Empirical Studies for Whether

Reform Should be Undertaken . . . . . . . . . . . . . . . . . . . 48

III. THE CAPACITY OF ISSUER CHOICE TO ACCOMMODATE DIFFERENCES

AMONG U.S. ISSUERS IN THEIR SOCIALLY OPTIMAL LEVELS OF

DISCLOSURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49A. The Range of Regimes Offered U.S. Issuers under Regulatory

Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501. The Need to Establish that Regulatory Competition

will Lead to an Appropriately Differentiated Setof Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

2. Choi and Guzman’s Approach . . . . . . . . . . . . . . . . 523. An Alternative Approach . . . . . . . . . . . . . . . . . . . . 544. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

B. Even if a Differentiated Set of Regimes does Develop, IssuerChoice’s Capacity to Customize is not Worth its Biastoward Underdisclosure . . . . . . . . . . . . . . . . . . . . . . . . 57

C. The Possibility of Providing Different Rules for DifferentIssuers Within a Mandatory Disclosure Regime . . . . . . 60

IV. THE COSTS OF TRANSITION TO ISSUER CHOICE . . . . . . . . . . . . . . . . 61A. Romano’s Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61B. Choi and Guzman’s Approach . . . . . . . . . . . . . . . . . . . . . . 63

V. FOREIGN ISSUERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

VI. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

*. Professor of Law, University of Michigan. B.A. 1968, J.D. 1971, Ph.D. (Economics)1980, Yale. The author wishes to express his appreciation for the helpful comments onearlier drafts of this article to Professors Lucian Bebchuk, John Beckerman, Omri Ben-Shahar, Nicholas Georgakopoulos, Peter Hammer, Louis Kaplow, Kyle Logue, RonaldMann, Mark West, Michelle White [others], Dean Joel Seligman and to participants at theHarvard Law and Economics Workshop and the University of Michigan Law andEconomics Workshop. Financial support for this project was provided by the Cook Fundof the University of Michigan. 1. Roberta Romano, Empowering Investors: A Market Approach to SecuritiesRegulation, 107 YALE L.J. 2359 (1998). 2. Id. at 2362. 3. Stephen J. Choi and Andrew Guzman, The Dangerous Extraterri tor ia l i ty o fAmerican Securities Law, 17 NW.J.INT’L L. & BUS. 207, 231-232 (1996) [hereinafter, Choi &Guzman, Dangerous Extraterritoriality]; Stephen J. Choi and Andrew Guzman, PortableReciprocity: Rethinking the International Reach of Securities Regulation 71 S. CAL. L.REV. 903 (1998) [hereinafter, Choi & Guzman, Portable Reciprocity]. 4. See, e.g., The Market for Regulation, ECONOMIST 82 (March 7, 1998). Also, on June12, 1998, the American Enterprise Institute held a seminar in Washington, D.C., entitled “ASecurities Market Approach to Securities Regulation” featuring Professor Romano andseveral discussants.

5/14/99 draft-e

RETAINING MANDATORY SECURITIES DISCLOSURE:WHY ISSUER CHOICE IS NOT INVESTOR EMPOWERMENT

Merritt B. Fox*

Professor Roberta Romano proposed recently that the mandatorysystem of federal securities law be replaced by a system of issuer choice.1 AnyU.S. issuer desiring not to be bound by the existing federal regime would beallowed to select instead the securities law regime of any of the fifty states orany foreign country. This reform, she argues, would “empower investors.” Itwould eliminate the federal government’s current “regulatory monopoly,” whereofficials make the rules to best satisfy their own agendas, and create in its placea market with multiple jurisdictions competing to offer issuers the regulations thatmaximize share value.2 Professors Stephen Choi and Andrew Guzman haveproposed essentially the same reform.3 To them, however, the primaryadvantage of issuer choice is that it would enable each issuer to choose fromamong a range of regimes the one best suited to the issuer’s particular securitiesregulation needs. As a result of these proposals, issuer choice is beginning toattract serious attention in policy circles.4

RETAINING MANDATORY DISCLOSURE 2

5. In their articles proposing issuer choice, Romano and Choi and Guzman also devotethe largest portion of their discussion to matters involving disclosure regulation. The othermajor component of security regulation concerns fraud: making false or misleadingstatements in connection with a securities transaction or trading on the basis of insideinformation. Since both of these activities, like disclosure, involve the quality ofinformation available to the investing public, many of the arguments against issuerschoosing their own disclosure regimes translate readily into arguments against issuerschoosing their own antifraud regimes as well. In at least one regard, the case against issuerchoice in antifraud regulation is even stronger than the case against it in disclosureregulation. This is because antifraud regulation applies to face-to-face transactions as wellas to impersonal market ones. In market transactions, the efficient market hypothesissuggests that the effect on share value arising from the issuer’s choice of regulatoryregime will be reflected in the market price even if the parties to the transaction are unawareof the choice or its implications. Thus the particular parties to the transaction are, throughlower price, compensated ex ante for any inadequacies in the chosen regime whatevernegative effects of these inadequacies may have on the economy as a whole. See I.C.3infra. In face-to-face transactions where one or both parties are unaware of the issuer’schoice of regime or its implications, there is no assurance of such ex ante compensation. 6. The debate began with empirical work by certain economists purporting to showthere is no benefit from the current mandatory U.S. disclosure regime. See, e.g., GeorgeStigler, Public Regulation of the Securities Markets, 37 J. BUS. 117, 122-124 (1964); GeorgeBenston, Required Disclosure and the Stock Market: An Evaluation of the SecuritiesExchange Act of 1934, 63 AM. ECON. REV. 132 (1973). This work and the work of othereconomists who come to opposing conclusions is discussed infra in Part II. Signaling –

This Article argues that despite these apparent attractions, we shouldreject issuer choice and retain the current mandatory system. My focus is on theimpact of issuer choice on disclosure regulation, the most important componentof securities law subject to the proposed reform.5 Giving U.S. issuers the rightto choose their disclosure regimes would likely decrease, not increase, U.S.economic welfare. For each U.S. issuer, there is a socially optimal level ofdisclosure. More information about the issuer and the resulting increase in itsshare price accuracy produces social benefits in the form of improved selectionof new investment projects, improved managerial performance and reducedinvestor risk. However, more information entails additional social costs as well.The issuer’s socially optimal level of disclosure is where the marginal socialbenefits equal the marginal social costs. Issuer choice would lead U.S. issuersto disclose at a level significantly below this social optimum.

To readers steeped in the history of corporate and securities lawscholarship, argument about mandatory disclosure will have a familiar, if distant,ring. While most commentators took mandatory disclosure’s desirability as agiven for the first three decades following passage of the Securities Act of 1993(“Securities Act”) and the Securities Exchange Act of 1934 (“Exchange Act”),a heated debate broke out starting in the mid-1960s.6 Opponents argued that


the idea that issuers with good news will want to disclose it and that the market will inferfrom the silence of the rest that they do not have good news – added a theoreticalcomponent to the case against mandatory disclosure. See Steven A. Ross, DisclosureRegulation in Financial Markets: Implications of Modern Finance Theory and SignalingTheory, in ISSUES IN FINANCIAL REGULATION (Franklin Edwards, ed. 1979). Signalling theoryis discussed infra in I.C. 7. John C. Coffee, Jr., Market Failure and the Economic Case for a MandatoryDisclosure System, 70 VA. L. REV. 717 (1984); Frank Easterbrook & Daniel Fischel,Mandatory Disclosure and the Protection of Investors, 70 VA. L. REV. 669, 684-85 (1984).The only prominent dissenter is Jonathan Macey, who argues: “[a]s markets have becomemore efficient, society’s need to devote resources to support a statutory regime ofmandatory disclosure designed and enforced by the SEC has disappeared. Anyinformation that was supplied by the force of law now is supplied by the marketplace.”Jonathan R. Macey, Administrative Agency Obsolescence and Interest Group Formation:A Case Study of the SEC at Sixty, 15 CARDOZO L. REV. 909, 928 (1994). This dissent isquestionable, however, since evidence that markets efficiently impound the informationthat issuers choose to release in no way shows that issuers will choose to release as muchinformation as is socially optimal. 8. It is quite possible that issuer choice would grant issuers as much freedom to choosetheir level of disclosure as no regulation at all. This would be the case if each jurisdiction,in its effort to attract issuers, chose to appeal to the preferences of a particular niche ofissuers. Under these circumstances, if a significant number of issuers wanted to discloseat any given level of disclosure (including disclosing essentially nothing), they would befree to do so because one or more jurisdictions would design their regimes to meet thisdemand. It is also possible, however, that each jurisdiction would try to maximize thenumber of issuers utilizing its regime by appealing to the broadest segment of the market,i.e., by choosing the level that minimizes the average distance between its requirements

market forces alone could provide sufficient incentives for issuers to disclose attheir socially optimal levels. By the mid-1980s, however, this debate had largelydied out. A rough consensus returned, with even most economics-oriented legalacademics, ranging from Professor John Coffee to Professors FrankEasterbrook and Daniel Fischel, concluding that on balance mandatorydisclosure should be retained.7

The proponents of issuer choice are clearly challenging this consensus.Admittedly, they offer a different alternative to mandatory disclosure from theone offered by its earlier opponents. Romano and Choi & Guzman wouldrequire each issuer to be bound by some disclosure regime, but allow it tochoose which one. The earlier opponents would have allowed issuers to bebound by no regime at all. This new alternative to mandatory disclosure,however, shares in common with the old alternative at least one core feature:they each grant issuers substantial freedom to choose their own disclosure levels.It is concern with how issuers would use such freedom that is at the heart of theprevailing consensus for retaining mandatory disclosure.8


and the preferences of each of the world’s issuers. While this would result in alljurisdictions offering regimes requiring the same level of disclosure, the level chosen wouldstill raise the same kind of concerns that generated the prevailing consensus againstswitching from mandatory disclosure to no regulation. This is because the level ofdisclosure would still be determined by preferences of issuers. These points are discussedin detail in Part III infra. 9. This is not to say that there are not differences that make issuer choice superior tototal abandonment of regulation–there are–only that issuer choice and total abandonmentof regulation share in common features that make each of them inferior to mandatorydisclosure. Compared to total abandonment of regulation, issuer choice has at least twoadvantages. It permits an issuer, at the time of a new issue of securities, to provideassurances to its investors that if its disclosures do not turn out to be a true and completeset of answers to a set of questions provided by the chosen regime, the issuer will facegovernmental sanctions. Issuer choice also permits the issuer to give a commitment,backed by governmental enforcement and sanctions, to providing an ongoing stream ofperiodic information at the level required by the chosen regime. In each case, however,since the issuer is free to choose its regime, it is free to choose, respectively, whatinformation has government sanction backed assurances of truthfulness and what level ofinformation it is committing to provide periodically.

This Article advances the reopened debate in two ways. First, itdemonstrates that the proponents of issuer choice have not effectively counteredthe arguments that have formed the basis of the prevailing consensus for retainingmandatory disclosure. While this consensus was formed when the alternative tomandatory disclosure was total abandonment of regulation, the proponents ofissuer choice have not shown how the arguments that form the basis of thisconsensus have any less force when the alternative is issuer choice.9 Nor havethey offered persuasive, more general rebuttals to these arguments. Second, thisArticle identifies several new and important arguments in favor of retainingmandatory disclosure.

The stakes in getting the right answer to this debate are potentially large.Securities markets are information driven and issuer choice would affect howmuch issuers disclose. Thus the decision whether to retain mandatory disclosureor switch to issuer choice may significantly affect how well our capital marketsperform their basic functions.

Part I of this Article considers the kind of disclosure regime that issuersare likely to adopt under issuer choice. An issuer’s managers, not its investors,will in the first instance make this choice. I will show that disclosure’s costs tothese managers are greater than its social costs and disclosure’s benefits to themare less than its social benefits. Each issuer will accordingly choose a regimerequiring significantly less disclosure than is socially optimal. Part II evaluates theexisting empirical literature bearing on the question of whether the current systemof mandatory disclosure enhances or diminishes social welfare. I find the results


10. Issuer choice will create an incentive for each jurisdiction to shape its requirementsto attract as many issuers as possible. Romano sees this as a virtue. In her eyes, it wouldbe a useful antidote to the tendency of the SEC, as a “monopoly” regulator, to adopt therules that suit its own agenda rather than the rules that would maximize share value.Romano, supra note 1, at 2362. For us to be assured that competition will in fact moverequirements in that direction, however, we would need to assume that when faced with a

of these empirical studies to be inconclusive. I also show that the question willprobably never be empirically resolved one way or the other. Showing thisreveals how inappropriate it is for the proponents of issuer choice to try to putthe empirical burden on advocates of retaining the current system, particularlygiven the strong theoretical argument that issuer choice would lead tounderdisclosure. Part III appraises the argument advanced by Professors Choiand Guzman that issuer choice would better accommodate differences amongU.S. issuers in their optimal levels of disclosure. I find that in practice issuerchoice would be unlikely to realize this hoped-for result. Part IV considers thetransition costs of adopting issuer choice. Part V considers the argument thatissuer choice improves capital mobility and reduces costs by permitting foreignissuers to choose their own regime when their shares are offered or traded in theU.S. market. I argue that the same advantages can be obtained with far fewerproblems by maintaining the current mandatory U.S. regime but redirecting itsreach so that it applies to U.S. issuers only. Part VI concludes.

The new round of debate concerning mandatory disclosure hassharpened our understanding of the role of disclosure regulation, and theproponents of issuer choice deserve credit for initiating this process. But ideasthat are good for provoking debate are not necessarily good policy. A switchto issuer choice would result in issuers disclosing significantly less than is sociallyoptimal. This consequence is undesirable unless the existing mandatory regimeresults in issuer disclosure behavior that deviates even further from what issocially optimal. There is no reason to presume that it does, and to date theproponents of issuer choice have offered no sustained argument, theoretical orempirical, to the contrary.

I. THE NATURE OF THE DISCLOSURE REGIMES ISSUERS WILL CHOOSE

The obvious starting point for an inquiry into the social welfare effects ofadopting issuer choice is to ask what kind of disclosure regime each U.S. issuerwould select if given the choice. Fundamental to the stories of both Romano andChoi and Guzman is a belief that the issuer will choose, from the regimesavailable, the one requiring it to disclose at the most socially beneficial level.10


choice, an issuer will choose the more socially beneficial regime over the less beneficialone.

Choi and Guzman see issuer choice and its ensuing jurisdictional competition as makingavailable a range of regimes, each with its own level of required disclosure. Each issuer,rather than being forced to comply with a purported “one-size-fits-all mandatory regime,can instead be matched with the regime most suitable for it from a social welfare point ofview. Again, there is no assurance that such a matching will occur unless we assume thatwhen faced with a range of choice, the issuer will choose the most socially beneficialregime available. 11. Professor Romano contemplates a system in which, for any given issuer, onesovereign – the U.S. federal government, a state, or a foreign country – would havejurisdiction over all transactions in the issuer’s securities. Romano, supra note 1, at 2362.She contemplates two possible ways that this sovereign could be chosen. One is that itwould be provided for by the firm’s founders in its articles of incorporation and the otheris that the founder’s choice of state of incorporation be deemed the jurisdiction governingsecurities transactions as well. Id. at 2408-10. Under either scheme, management couldsubsequently change the jurisdiction governing securities transactions, but to do so itwould need to obtain the approval of the issuer’s shareholders. Id. at 2415-16. Romanodoes not explicitly address the question of how existing, publicly traded issuers would betreated at the time that issuer choice is adopted. She presumably intends, however, thatthey initially would continue to be under the U.S. federal regime until and unlessmanagement decides to switch to some other jurisdiction and obtains the necessaryshareholder approval. For a discussion of why I believe that shareholder approval is nota meaningful check on a management decision to make such a switch, see note 53 infra.

Professors Choi and Guzman suggest that at the time of each new issue of securities, anissuer’s management would choose the jurisdiction whose rules would govern transactionsin securities of that particular issue. Choi & Guzman, Portable Reciprocity, supra note 3,at 922. They propose no exit mechanism once that choice is made.

In fact, each issuer will prefer a regime that requires it to disclose substantiallyless.

A. The Persons Choosing an Issuer’s Regime: Private versus SocialOptimality.

Under the issuer choice approach, an issuer’s managers are theindividuals who actually choose the issuer’s disclosure regime.11 Thus, it is theirpreferences and incentive structure on which we must focus. If the choice isbeing made at the time of an initial public offering, the managers will themselvesbe, or be allied with, owners of a substantial portion of the issuer’s shares. If thechoice is being made when the corporation is already a large, establishedpublicly traded corporation, the managers may own only a tiny percentage of theissuer’s shares.


12. This somewhat stylized model involves the issuer disclosure level being measuredcardinally and, as depicted in Figure 1, having associated with it a rising marginal costcurve and a falling marginal benefit curve. This is a reasonable depiction of what likelyhappens. The managers first have the issuer release the piece of information that is mostbeneficial to them relative to its cost, then the piece of information that is next mostbeneficial relative to its cost, and so on. This corresponds to a situation where marginalprivate benefit is decreasing and marginal cost is increasing. 13. Since there is no reason to believe that the difference between the marginal socialbenefit and marginal private benefit, or between the marginal social cost and marginalprivate cost, would increase substantially as an issuer discloses more, the marginal socialbenefit should, as depicted in Figure 1, also be decreasing and the marginal social cost alsoincreasing. See supra note 12.

Assume that the managers have a range of regimes from which tochoose, each requiring a different level of disclosure. A regime requiring greaterdisclosure involves both added costs to the managers and added benefits forthem. The marginal cost of additional disclosure tends to rise; the marginalbenefit tends to decline.12 The managers will choose the regime that requires theissuer to disclose closest to the level at which the marginal increase in cost to themanagers (the private marginal cost) equals the marginal increase in benefit tothem (the private marginal benefit). This will be below the issuer’s sociallyoptimal level of disclosure because, as shown below, over the whole range oflevels at which an issuer could disclose, the social marginal cost of an issuer’sdisclosure is below the private marginal cost to its managers, and the socialmarginal benefit of its disclosure is above the private marginal benefit to theseindividuals.13


14. I have considered this point in more detail elsewhere. See Merritt B. Fox, SecuritiesDisclosure in a Globalizing Market: Who Should Regulate Whom, 95 M ICH. L. REV. 2498,2537-39 (1997). [hereinafter, Fox, Disclosure in a Globalizing Market]. See also EdmundW. Kitch, The Theory and Practice of Securities Disclosure, 61 BROOK. L. REV. 763 (1995);Coffee, supra note 7; Easterbrook & Fischel, supra note 7, at 684-685; Lucian A. Bebchuk,Federalism and the Corporation: The Desirable Limits on State Competition in

B. Costs.

1. The Divergence of Private from Social Costs

For each individual U.S. issuer, a disclosure involves two differentkinds of costs, “operational” costs and “interfirm” costs. Operational costs arethe out-of-pocket expenses and the diversions of management and staff time thatissuers incur to provide the information. Interfirm costs arise from the fact thatthe information provided can put the issuer at a disadvantage relative to itscompetitors, major suppliers and major customers. Operational costs are costsboth to the individual firm and to society as a whole. Interfirm costs are costsonly to the individual firm. They are not social costs because the interfirmdisadvantages to the issuer from the disclosure are counterbalanced by theadvantages it confers on the other firms. Thus, at all levels of disclosure, anissuer’s private marginal costs will exceed its social marginal cost by an amountequal to these interfirm costs.14 Even managers who completely identify with


Corporate Law, 105 HARV. L. REV. 1435, 1490-91 (1992). 15. Romano, supra note 1, at 2426. Professors Choi and Guzman acknowledge theexistence of interfirm costs. Stephen J. Choi & Andrew Guzman, National Laws,International Money: Regulation in a Global Capital Market, 65 FORDHAM LAW REVIEW

1855, 1875 (1997) [hereinafter, Choi & Guzman, National Laws]. They do not, however,attempt to confront the negative implications of this market failure for the reform theyadvocate. 16. Romano, supra note 1, at 2368. 17. Roland A. Dye, Mandatory Versus Voluntary Disclosures: The Cases of Financialand Real Externalities, 65 ACCT REV. 1 (1990). 18. Dye contemplates a situation in which an entrepreneur, in advance of a sale ofsecurities to outside investors, commits to providing information that will reduce by somegiven amount the uncertainty associated with his issuer’s future cash flow to shareholders.Dye compares the amount of disclosure that entrepreneurs can be expected to commitvoluntarily to provide with the amount that is socially optimal. Dye considers the effects

existing shareholders – managers who seek to maximize share value so that coststo the shareholders are equivalent to costs to them – would therefore choose aregime with a disclosure level below the social optimum.

2. Romano’s Dismissal of the Importance of the Private/Social CostDivergence

This divergence of private from social costs means that issuer choice willlead to market failure. It thus presents a serious problem for the proponents ofissuer choice. Professor Romano, however, dismisses the importance of thedivergence, suggesting that it is “a tenuous rationale for securities regulation.”15

She bases this dismissal on four lines of attack, each of which is unconvincing.a. The claim that as a matter of theory the divergence of

disclosure’s social from private cost does not necessarily call forregulation. Romano’s first line of attack is to suggest that even if this divergenceexists, the theoretical case for mandatory disclosure is not as clear cut as Isuggest.

It can be shown analytically ... that even in the case of such third-party externalities [information that would hurt the issuer byhelping its competitors], mandatory disclosure is not alwaysoptimal compared to voluntary disclosure, and it would in alllikelihood be extremely difficult for a regulator to determine whenmandatory disclosure is optimal.16

She does not explain why this is the case, but instead simply cites an article byRonald Dye.17 Dye’s model, however, applies only in such a restricted range ofcircumstances as to be essentially useless as a guide to policy.18


of two kinds of third-party effects that can arise from an issuer’s disclosure. The first kindis “financial externalities,” wherein the issuer’s disclosure reduces uncertainty concerningthe future cash flows to shareholders of one or more other issuers, but has no effect on thecash flows of other issuers. Id. at 3. This first kind of third-party effect relates todisclosure’s “public goods” aspect. See, Romano, supra note 1, at 2367 and infra I.C.2(where it is discussed as one of the factors that causes the social marginal benefit fromdisclosure to be greater than its private marginal benefit). The second kind of third partyeffect is “real externalities,” wherein the issuer’s disclosure affects the cash flows of oneor more other issuers. Dye, supra note 17, at 3. It is this second kind of third party effectthat is the subject of the discussion above and to which Romano refers in the quotation.

Dye concludes that under some circumstances, the voluntarily committed level ofdisclosure will be at least as great as the socially optimal one, thereby rendering mandatorydisclosure unnecessary, while under other circumstances the voluntarily committed levelwill fall short of that goal. He suggests that where there are only financial externalities, thevoluntary level would equal the socially optimal one under a fairly wide range ofconditions, id. at 15, but that “where real externalities are present, optimal and equilibriumvoluntary disclosure tend to diverge.” Id. at 3. He also suggests that with realexternalities, the direction of the divergence – whether voluntary disclosure produces toomuch information or too little – is unclear in the absence of detailed information. Id. at 19.A number of factors, however, severely limit the generality of Dye’s conclusions andrender the model relatively useless as a guide to policy.

1. The empty set of information to which the model applies. Dye makes two criticalassumptions concerning the type of information to which his model applies. One is thatmanagement does not know the information that it is committing to provide at the time itmakes the commitment. The other is that the information is disclosed prior to sale of theshares. Id. at 3-4.

Disclosures made by real-world issuers will almost always fail to conform with one or theother of these assumptions. Before an issuer has public holders of its securities, it has noneed to release publicly any information about its financial condition and prospects. Whenthe issuer’s entrepreneurs decide to have a public sale of its securities, however, they arelikely to decide to release some such information just before the sale. Strictly speaking,there is no commitment associated with this release; nothing in advance of the releasebinds the entrepreneurs to make it. The released information may be accompanied,however, by a pledge that it represents a truthful set of answers to a recognized set ofquestions. One could characterize certified financials in this fashion, for example. But evenif Dye’s use of the term “commitment” is stretched to include such a pledge, thecommitment is made simultaneously with the release of the information, and so it involvesinformation known to the entrepreneurs. Thus it does not fit within Dye’s model becauseit violates the model’s first assumption.

Something else is likely to happen at the time of the public sale of securities, however.The entrepreneurs are likely to commit to providing additional information periodically inthe future. An example of such a commitment would be the decision to list on a stockexchange that requires such ongoing disclosure. Although an act of this sort doesrepresent a real commitment as the term is generally understood, it does not fit into Dye’smodel either. While the information ultimately disclosed pursuant to this commitment maynot be known by the entrepreneurs at the time they make the commitment, the disclosurewill occur after the sale of the securities. This violates the second assumption of Dye’s


model.Thus the model is not useful in comparing voluntary versus mandatory disclosure with

regard to either new issue disclosure (currently subject to the mandatory requirements ofthe Securities Act of 1933) or periodic disclosure (currently subject to the mandatoryrequirements of the Securities Exchange Act of 1934).

2. The exclusion of share sales to fund new investment. The only kind of transactioncontemplated by the model is the sale by an entrepreneur of his own shares. The issuerinvolved has already raised its needed capital and made its investment. The sale occurssimply because the entrepreneur wishes to consume before the investment produces itscash flow. Id. at 3-5. Thus the model does not reach disclosure associated with a sharesale by an issuer to raise capital to fund a new project, a more common kind of transactionthat is probably more vital to the economy.

3. The misspecification of disclosure’s costs and benefits. The only function fordisclosure identified by the model is its potential for altering the allocation of risk betweenselling entrepreneurs and purchasing outside investors. According to the model,disclosure that conforms to the models’ two assumptions will bring the price that investorswill pay for their shares closer, one way or the other, to the discounted present value of thecash flows that they in the end actually receive, thereby reducing the riskiness of theirdecision to invest. The commitment to provide such disclosure creates a correspondinguncertainty as to the price that the entrepreneur will receive: he knows the disclosure willaffect the price, but he does not know which way because he does not know its content.Thus, in terms of risk, disclosure benefits risk averse investors and harms risk averseentrepreneurs.

This result, however, is an artifact of the model’s particular features discussed above,whereby entrepreneurs are the sellers of the shares and commit to provide, before the sale,information not currently known to them. Neither new issue disclosure nor periodicdisclosure of the kinds that issuers provide in the real world will have this risk reallocationeffect. Consider first new issue disclosure. For the entrepreneur, the information disclosedis in reality already known to him. Thus the “commitment” associated with its provision,rather than creating uncertainty as to the price the entrepreneur will receive, has apredictable impact on that price. The information will not, however, be known to theinvestors prior to disclosure. Its receipt will reduce each investor’s uncertainty as to thereturn on an investment in the issuer’s shares. With disclosure, the price she pays will onaverage be closer, on one side or the other, to the cash flow she will actually receive in theend. This does not necessarily mean that it will reduce her exposure to risk, however.Disclosure by an individual firm concerning its financial condition and prospects relatesto uncertainties that are uncorrelated with uncertainty concerning the aggregate cash flowsgenerated by all the firms in the market. To say this in the language of the capital assetpricing model (CAPM), individual firm disclosure only reduces unsystematic risk. This canbe true even if the firm’s disclosure reduces uncertainty concerning the cash flows ofcertain other firms (for example, ones within the same industry), thereby creating financialexternalities. Thus if the investor is purchasing shares to be part of a fully diversifiedportfolio, the reduction in unsystematic risk will not reduce the riskiness of her overallportfolio, which is all that matters in terms of her exposure to risk. An investor purchasingthe shares to be part of a less than fully diversified portfolio would benefit from thedisclosure, but CAPM suggests that, unlike in Dye’s model, the entrepreneur will notreceive a higher price as a result of providing this benefit. See infra I.C.1.


Now consider the commitment to provide periodic disclosure. It likewise will notincrease risk for the entrepreneurs nor reduce it for investors. The commitment to providethe information will not create uncertainty concerning the price that the entrepreneur willreceive because the information will not be disclosed until after the sale and thereforecannot affect that price. For the same reason, it cannot bring the price that the investorpays closer to the cash flow she will ultimately receive. The fact that with periodicdisclosure the investor at some point after the sale will have a more accurate perception ofthese cash flows is irrelevant to the riskiness of her investment decision, which isdetermined by how likely it is (based on what is known at the time of the sale) that thesecash flows will deviate substantially one way or the other from the price paid for the shares.

Dye’s model thus seriously misspecifies disclosure’s potential costs and benefits, bothprivate and public. It focuses on a factor on which real world disclosure is likely to havelittle or no impact – the allocation of risk between entrepreneurs and investors – whileignoring its important effects, discussed below, on capital allocation and the agency costsof management. See infra I.C.2 and I.C.3. 19. In his analysis of real externalities, Dye finds that if an issuer experiences negativeprivate returns from additional disclosures but they lead to positive market-wide returns,the voluntary level of disclosure will be below what is socially optimal. Dye, supra note17, at 19. He observes that the opposite would be true if the issuer experiences positiveprivate returns from additional disclosure but the market experiences negative returns fromit. Id. He suggests that plausible scenarios can be constructed either way and that

without possessing detailed a priori knowledge about the relation between private andmarket wide returns to additional disclosures for each firm, it is difficult to surmisewhether [socially optimal] mandatory disclosures will exceed voluntary disclosures.

Id.The signs of the private and market returns to disclosure are not in fact as indeterminate

as Dye states. Specifically, I suggest above that there is a wide range of items ofinformation for which one of the private costs of disclosing the item is its harm to thedisclosing issuer’s competitive or bargaining position vis a vis other firms. This particularcost is counterbalanced by a real externality, the corresponding gains enjoyed by the otherfirms. Thus their combined effect on market-wide returns is a wash. The other privatecosts associated with the item’s disclosure have no such corresponding positive externaleffects, nor do the private benefits associated with its disclosure have any correspondingnegative external effects. Consequently, the social value of the disclosure – what Dyewould term its effect on market-wide returns – is at least as great as these private benefitsminus these other private costs. For all items near the socially optimal level of disclosure,the effect of their disclosure on private returns (private benefits minus total private costs,including the ones related to the issuer’s competitive and bargaining positions) will benegative, while the effect of their disclosure on market wide returns will be positive. Firmswill thus not voluntarily disclose as much as is socially optimal.

More generally, it is hard to imagine how, as an ordinary matter, a disclosure by any oneissuer would have a negative effect on the aggregate cash flows of all the other firms in theeconomy. If such a disclosure affects the aggregate cash flow of these other firms at all(i.e., if a real externality is present), it will be in the positive direction. This is becauseadditional information allows the other firms to better predict the consequences of their

Moreover, when reasonable estimates of the model’s parameters are utilized, hismodel in fact supports the case for mandatory disclosure.19


actions. Thus, contrary to Dye, there is little possibility that the private returns todisclosure are positive and the market-wide ones negative. Dye’s own model, therefore,suggests that real externalities create a divergence between the voluntary level ofdisclosure and the optimal one and that the direction is toward too little voluntarydisclosure. 20. Romano, supra note 1, at 2368. 21. The analysis that follows in the text also applies to the choice of a regime to governan issuer’s ongoing periodic disclosures if it is made at the time of the issuer’s initial publicoffering. 22. In a public market for securities, there is no opportunity for bargaining between theholders of B shares and the managers of A whereby the purchase by the B shareholderscould be conditioned upon the A managers choosing the high disclosure regime.

b. The claim that investors holding diversified portfolios solve theproblem. Romano next argues that the externality producing the divergencebetween social and private costs does not require mandatory disclosure because“the majority of investors hold portfolios ... and therefore, unlike the issuer, theywill internalize the externality if they make the disclosure decision.”20 Thisargument too does not stand scrutiny.

To start, the argument has no applicability to new issue disclosure, whichis a large part of the whole regulatory scheme.21 To see this, consider firm A,which is about to engage in an initial public offering and whose managers aredeciding whether to choose a high disclosure regime or a low disclosure one,and firm B, which is a competitor of A. The high disclosure regime wouldrequire a release of information that, by enhancing B’s ability to compete with A,would benefit B but harm A. The only way that this externality could beinternalized would be if the sale price for A’s shares would somehow beenhanced because of the benefit that the choice of the high disclosure regimewould confer on B. It will not be. Potential purchasers of the A shares fall intotwo groups: those who currently hold B shares and those who do not. Thosewho do not obviously have no reason to pay more for the A shares because ofthe benefit to B from the A managers choosing the high disclosure regime.Potential purchasers who do hold B shares have no reason to pay more either.Whether they buy or not, they will benefit from A choosing the high disclosureregime and suffer from A choosing the low disclosure one.22

Romano’s argument is unpersuasive with respect to periodic disclosureas well. This time, consider firm C, an existing publicly held issuer with all of itsshares already outstanding and whose managers are deciding whether to choosea regime requiring a high level of ongoing periodic disclosure or one requiring alow level of such disclosure, and firm D, a competitor of C. The high disclosureregime would require release of information that would, by enhancing D’s ability


23. Holding twenty properly chosen stocks achieves 95% of the risk reduction thatwould be obtained from holding the whole market portfolio and holding 100 such stocksachieves 99% of such reduction. RONALD GILSON AND BERNARD BLACK, THE LAW AND

FINANCE OF CORPORATE ACQUISITIONS 92 (2d. ed. 1995). Only an investor that holds aportfolio consisting of the same percentage of all the stocks available in the market – anindex fund – would privately experience costs and benefits from the disclosure that parallelits social costs and benefits. Yet index funds are notoriously passive concerning corporategovernance. See John C. Coffee, Jr., Liquidity versus Control: The Institutional Investoras Corporate Monitor, 91 COLUM. L. REV. 1277 (1991). 24. In addition, assuming that fewer than all of C’s shareholders hold shares of D, itwould be a breach of fiduciary duty on the part of the managers of C to account for thederivative benefits to the shareholders who do hold D shares when the managers make thedecision whether to adopt the higher disclosure regime.

The more reliable devices for helping align management interests with those ofshareholders – the hostile takeover threat and share price based managerial compensation– will be of no help in pushing C’s managers to account for the derivative benefit from thechoice of the higher disclosure regime that would accrue to C shareholders who hold Dshares as well. These devices depend on share prices. A firm’s share price is a functionof the market’s perception of the prospective cash flows that managers can produce, notof the good things it does for another firm. A shareholder is extremely unlikely to add orsubtract C shares from her portfolio based on whether management has or has not madea disclosure benefitting D.

to compete with C, benefit firm D but harm C. The fact that most Cshareholders hold diversified portfolios will not eliminate this externality. Tostart, few of these shareholders are likely to have portfolios that include sharesof D. This is because of the basic lesson of corporate finance that an investorcan enjoy most of the benefits of diversification by holding only a few dozenstocks out of the many thousands available.23 Thus, in all likelihood, the holdersof the majority of C’s shares would have no D shares and hence nocompensating interest that would counterbalance the competitive disadvantageof the high disclosure regime.

Even in the unusual situation where the holders of a majority of C’sshares do hold D shares as well, they are unlikely, when they vote their shares,to take account of the benefits from C choosing the higher disclosure regime.Their focus is going to be primarily on the cash flows that the C managers canproduce. The vast information asymmetries that exist between shareholders andthe managers, combined with well known collective action problems, makeshareholder influence tenuous enough without the introduction of additionalcriteria for judging management such as the effect on D of C management’schoice of disclosure regime.24

If Romano’s argument were correct, the fact that shareholders holddiversified portfolios would make equally unnecessary having legal sanctions in


25. Romano, supra note 1, at 2380. 26. One of the major types of information required by the Exchange Act that was notprovided by many corporations prior to its enactment in 1934 is sales and cost of sales.See II.A1 infra. Kaplan and Reaugh, in an article surveying, among other things, theannual reports of major corporations in 1930, after finding that 47% omitted their net salesfigures, stated:

The usual reason given for refusing to disclose sales and cost of sales figures is thecreation of consumer resistance where the gross profit margin is wide and that itspublication invites competition or gives an advantage to existing competitors, especiallywhere the competitors’ figures remain undisclosed.

Under the Securities Exchange Act ... corporations ... are required to list net sales intheir ... periodic reports.

Maurice C. Kaplan & Daniel M. Reaugh, Accounting, Reports to Stockholders, and theSEC, 48 YALE L.J. 935, 946 (1939). See also TWENTIETH CENTURY FUND, INC ., THE

SECURITIES MARKETS 580-81 (1935). 27. See, e.g., Regulation S-K, 17 CFR §229.101(b) (1998); §229.10(a)(2)(iii)(B)(2);§229.101(c)(1)(xi); §229.303(a)(3)(ii); §229.303(a)(1); §229.101(c)(1)(viii); §229.101(c) (1)(vi);§229.101(c)(1)(iii).

There is also anecdotal evidence that supports the existence of significant interfirmcosts. My personal experience in legal practice, and that of other securities practitioners

response to any type of corporate behavior that negatively affects othercorporations, such as patent infringement or even breach of contract. Thisobviously would make no sense.

c. The claim that proponents of mandatory disclosure have notestablished the existence of any kind of disclosure having interfirm costs.Romano also suggests that persons who argue that the divergence of privatefrom social costs justifies mandatory disclosure “have not suggested whatinformation requirements the rationale justifies, let alone whether that informationis the focus of SEC disclosure requirements.”25 It seems self-evident, however,that almost all potential corporate disclosures have interfirm costs associatedwith them. This was one of the main factors that drove adoption of mandatorydisclosure in the first place.26 That this in fact is the focus of SEC requirementscan be confirmed by even a quick review of SEC Regulation S-K, whichprovides the questions that are incorporated by reference into its forms 10-K(for periodic disclosure) and S-1 (for initial public offering disclosure).Regulation S-K calls for a wide variety of information the disclosure of which onthe one hand would be useful for predicting an issuer’s future cash flows, but onthe other could seriously hurt the issuer through the advantages it confers onother firms. Examples include profits and sales of each significant individual lineof business conducted by the issuer, future capital spending plans, research anddevelopment spending, cost ratios, liquidity constraints, and information onbacklogs, inventories and sources of supply.27


with whom I have talked, is that at the margin, where it is not absolutely clear whether thesecurities laws require disclosure, issuers often resist providing it, giving as their mostfrequent reason their fear that the information will be used by their competitors. 28. Romano, supra note 1, at 2380-81. 29. Id. 30. See infra I.C.3. 31. The relationship between more information and price accuracy is discussed in I.C.3,infra. The social gains from improved price accuracy are discussed in I.C., infra. 32. Bipin B. Ajinkya, An Empirical Evaluation of Line of Business Reporting, 18 J.ACCT . RES. 343 (1980); Rosanne M. Mohr, The Segmental Reporting Issue: A Review ofEmpirical Research, 2 J. ACCT . LIT. 39, 56-57 (1983) (survey article collecting studies). 33. Mohr, supra note 32, at 42-55 (survey article collecting studies). For a review of theliterature concerning the proposition that mandatory disclosure as a general matterenhances price accuracy, see Fox, Disclosure in a Globalizing Market, supra note 14 at2540-2541, and II.A.1, infra.

d. The claim that issuers will not comply. Finally, Professor Romanoargues that the divergence of private from social cost cannot justify a mandatorydisclosure statute because wherever such a divergence is present, issuers willresist providing the required disclosures. Hence the statute is not implementable.Romano gives as her example the SEC requirement, imposed in 1969, thatissuers provide line-of-business (LOB) reporting.28 She cites a number ofstudies that she concludes prove the ineffectiveness of LOB reporting rules. Shebases this conclusion on the fact that the studies show no statistically significantincrease in actual returns after imposition of the requirements.29 These results arenot surprising, however, because LOB disclosure is going to reveal as manysituations where the market would otherwise overvalue shares as where it wouldotherwise undervalue them.30

Contrary to her conclusion, the studies cited by Romano in fact provideaffirmative evidence that the LOB reporting rules have resulted in issuersproviding meaningful new information to the market. If there is meaningful newinformation in the market, share prices become more accurate.31 The studiesshe cites demonstrate that LOB disclosure indeed did lead to an improvementin price accuracy. They contain direct evidence of improved price accuracy inthe form of results showing reduced total variance and price dispersion.32 Theycontain indirect evidence as well in the form of results showing increasedaccuracy of analyst forecasts, which in turn should lead to improved share priceaccuracy.33

3. The Further Divergence of Private from Social Cost Because ofManagement’s Agency Relationship with the Shareholders.


34. In the preceding discussion of the other source of divergence – interfirm costs – itwas assumed that managers completely identify with existing shareholders. This permittedthe costs of disclosure to the firm as a whole to be treated as equivalent to disclosure’scosts to the managers, the actual disclosure decision-makers. This is reasonable since firmcosts are costs to the managers as well: disclosure, by damaging the firm, derivativelydamages the managers also and does not provide them with any compensating gains. Theassumption, however, understates the total cost of disclosure to the managers since itignores the fact that disclosure can harm their position in their agency relationship withshareholders. 35. See infra I.C.3. This increased managerial discipline would have a positive influenceon share price from which the managers may gain a derivative benefit. The size of thisbenefit, however, would in many cases not be sufficient to cancel out the personal cost tothem of having to work under greater discipline. Id. 36. See Romano, supra note 1, at 2366-67. Choi and Guzman appear somewhatinconsistent on this question. On the one hand, in an earlier work where they propose aterritorial approach to securities choice of law in order to give issuers some degree ofchoice, they recognize a scenario – labeled by them a “race to the bottom” – by whichmanagers would choose to bind the issuer to provide a socially suboptimally low level ofdisclosure. In this scenario, managers gain from engaging in opportunistic behavior. Thefact that it is opportunistic implies that the joint wealth of the shareholders and themanagers would be greater if the managers did not engage in this behavior. Greaterdisclosure would make the opportunistic behavior impossible. A credible promise thatsuch greater disclosure is forthcoming would be reflected in increased proceeds from ashare issuance. As Choi and Guzman tell it, the managers’ benefits from the increasedproceeds are only “indirect”, however, and not as great as the managers’ gains from theopportunistic behavior. They therefore would choose a lower disclosure regime. Choi andGuzman, National Laws, supra note 15, at 1872. Thus, while increased disclosure wouldincrease the joint wealth position of the parties and represent a social gain, it would notoccur unless it was mandated because the managers do not feel its full social benefits and

The fact that an issuer’s managers are in an agency relationship with itsshareholders gives rise to an additional source of divergence between themanager’s private costs of disclosure and the social costs of that disclosure.34

Periodic disclosure increases the effectiveness of a number of devices–theshareholder vote, shareholder enforcement of management’s fiduciary duties,and the hostile takeover threat–that work to limit the ability of managers todeviate from acting in the shareholders’ best interests. This is a cost tomanagers, who would prefer to pursue their personal goals under as fewconstraints as possible.35

C. Benefits

Proponents of issuer choice rely on an unstated premise that the privatemarginal benefit associated with an issuer’s disclosure equals its social marginalbenefit.36 Such a premise involves two assumptions. The first is that the social


would choose a lower disclosure regime instead if given the choice.On the other hand, Choi and Guzman, later in the same article, state that a “separating

equilibrium” that would permit managers of such an issuer to choose a lower disclosureregime would be socially preferable to a regime that would require the issuer to disclose atthe higher level. Id. at 1877-78. It appears that at this later point they do not seem torecognize that they are advocating a reform that under these circumstances would lead toa decrease in social welfare. They do not address the question at all in their piecesproposing unfettered issuer choice. 37. See infra II.A.1. 38. According to the model, the market does not reward reductions in a share’sunsystematic risk because unsystematic risk can be eliminated entirely by holding a fullydiversified portfolio. The fact that less than fully diversified investors could protectthemselves in this fashion suggests that this gain from greater disclosure is a lesscompelling reason for having mandatory disclosure than the reasons discussed below.Nevertheless, since less than fully diversified investors exist and will continue to existwhatever the government does, the reduction in risk for these investors is a social gain.

benefits of the level chosen will be fully reflected in the issuer’s share price. Thesecond is that this share price improvement will be fully enjoyed by the personsmaking the choice–the issuer’s managers. The proponents of issuer choice donot spell out mechanisms that would generate either of these results. In fact,neither of these assumptions is correct much of the time. Thus the privatemarginal benefit associated with an issuer’s disclosure is likely to be below thesocial marginal benefit, adding to the extent of failure if the disclosure level isdetermined by market forces.

1. Reduction in investor risk.

When more information is available about an issuer, its share price islikely to be closer, on one side or the other, to actual value.37 This is a gain tothe less than fully diversified investor, because it reduces the risk of holding theissuer’s shares in her portfolio. Because the risk that is reduced is unsystematic,however, the issuer’s share price will not on average be any higher–thefundamental lesson of the capital asset pricing model.38 Thus there is nocorresponding reward to the issuer’s managers.

2. Improved project choice.

A second social benefit of disclosure is an improved choice amongproposed new investment projects in the economy. How this works can be seenmost easily in a simplified world in which each new investment project isundertaken by a new issuer that raises the necessary funds through an initial


39. Such a project has a negative net present value (evaluating its expected future cashflow on the basis of all available information including that available to each of theeconomy’s entrepreneurs, and using the expected return on the marginal project, adjustedfor risk, as the discount rate). Nevertheless, if the entrepreneurs go ahead, the sharesretained by the entrepreneurs have value because once the investment in the project ismade using the proceeds of the share sale, the project will produce an expected positivefuture cash flow. The retained shares represent a pro-rata claim on this expected positivecash flow.

public offering (“IPO”). An issuer’s managers will proceed with their proposedproject if and only if share price is high enough that the issuer can raise the cashnecessary to implement the project without having to offer all of its equity to thepublic. When this is the case, the managers can retain the rest of the equity forfree. The retained shares constitute an “entrepreneurial surplus” that gives thema pro rata claim on the expected positive cash flow generated by the project.Thus in this world, the private benefits from disclosure’s improvements in projectchoice would differ from the social benefits only to the extent that the socialbenefits of the chosen disclosure level are not fully reflected in price. To theextent they are, they will be directly enjoyed by the entrepreneurs making thedisclosure decision in the form of a larger entrepreneurial surplus.

Ideally, society would want to implement all proposed projects in rankorder of their risk-adjusted expected returns (based on all available informationincluding what is known by the managers proposing each project). The marginalproject that just exhausts society’s scarce savings for investment would set therisk-adjusted expected return on capital. Whichever issuer’s shares an investorpurchases, she would receive a risk-adjusted expected return just equal to thatof the marginal project.

This ideal will not be achieved in the real world because some of theinformation possessed by each proposed project’s entrepreneurs will not bepublic and hence not fully reflected in share price. Some projects inferior to theideal world’s marginal project will be implemented because their issuers’ shareprices are inaccurately high. Their entrepreneurs will gladly go ahead becausethey can raise all the cash necessary to implement their projects by offering thepublic less than all of the equity of their respective firms and keep the rest asentrepreneurial surplus.39 Likewise, some projects superior to the ideal world’smarginal project will not be implemented because their issuers’ share prices areinaccurately low. With such a project, a public offering of even all of the issuer’sequity would not produce sufficient cash to fund the project.

An increase by all issuers in their level of disclosure would increase theaccuracy of the price of every issuer’s shares. The resulting reduction in thenumber of misallocations would be a social benefit that, if greater than the social


40. By “unbiased,” I mean that the price is on average equal to the share’s actual value,i.e., what the future income stream accruing to the holder of the share – its dividends andother distributions – turns out to be, discounted to present value. Speculators – thepersons whose actions in the market set prices – assess what this future income stream willbe based not only on what information is available about the issuer but also on what theissuer does not disclose. The empirical literature testing the efficient market hypothesissuggests that the inferences that speculators draw from issuer disclosures are in factunbiased. Since there is no reason to believe that their inferences from issuer absences ofcomment are any more likely to be biased than their inferences from issuer disclosures, thisliterature suggests as well that the inferences they draw from issuer absences of commentare also unbiased. I discuss these points in considerably more detail elsewhere. See Fox,Disclosure in a Globalizing Market, supra note 14, at 2533-39.

costs involved, would represent an increase in social welfare (a “net” socialgain). The question for us, however, is what kind of incentives exist for themanagers of each individual issuer to produce the level of disclosure that isneeded to achieve this social gain. In essence, to what extent would any socialgain produced by a single issuer’s increased disclosure be reflected in its shareprice?

If we were to pick one issuer at random and command an increasein the amount of disclosure it provides, its managers would not on averageenjoy any perceptible benefit even if the cost to them of the disclosure werezero. This is a critical point that is left out of the existing literature debating thedesirability of mandatory disclosure. In an efficient market, the issuer’s shareprice without the increased disclosure would be an unbiased estimate of thefuture cash returns to the share’s holder discounted to present value and sowould the share price with the increased disclosure.40 Each price – the one inthe market with less disclosure and the one in the market with more disclosure– is generated by a probability function with a mean equal to the share’s actualvalue. Comparing market pricing with less versus more disclosure is analogousto taking different sample sizes from an urn containing many red and green ballsin order to estimate the ratio of the two in the urn. A small sample’s ratio, likea price in a market with low disclosure, will be an unbiased estimate of the actualratio in the urn. In other words, if many such small samples were taken, theiraverage ratio would equal the actual ratio in the urn, but any given such sampleis likely to be off, one way or the other. A large sample’s ratio, like a price ina market with high disclosure, will also be an unbiased estimate of the actual ratioin the urn. The only difference is that a large sample’s ratio will be moreaccurate in the sense that it will be closer, one way or the other, to the actual


41. The issuer’s increased disclosure will on average result in better capital allocation.This will increase the overall expected rate of return on capital in the economy. Thisincrease in the overall supply of expected future dollars would lower the rate at which thiscash flow (and that of all other investment opportunities) is discounted. In a largeeconomy that implements many proposed projects, however, the effect of this singleimprovement in capital allocation will have only a tiny effect on the discount rate. Thus theeffect on the price of the issuer’s shares would be imperceptible. See infra II.A.2.

The increased disclosure will also reduce the overall risk associated with the issuer’sshares but not in a way that will enhance share price. Since the information disclosed isfirm specific, all of the reduction will be in the share’s unsystematic risk. The market doesnot reward reductions in a share’s unsystematic risk with a higher price. See supra note38. 42. Signaling theory is the theory of self-induced disclosure in the context of an IPO.No one has offered any other plausible argument as to why voluntary disclosure alonecould be sufficient. The classic statement of signaling theory is found in Ross, supra note6. 43. See Easterbrook & Fischel, supra note 7, at 687-88. Commentators have also notedthat empirical reality does not conform with signaling theory’s prediction that voluntarydisclosure will result in the market being informed at the socially optimal level. See JoelSeligman, The Historical Need for a Mandatory Corporate Disclosure System, 9 J. CORP.L. 1, 7 n. 24 (1983); Coffee, supra note 7, at 745. Coffee points out that the market was not

ratio in the urn. Thus on average the issuer’s share price with high disclosure willbe the same as with low disclosure.41

If, however, we were to observe an issuer voluntarily picking itselfout to provide more disclosure than other issuers provide–to choose astricter regime on its own– we would on average expect the issuer’s shareprice to go up and its managers to enjoy a greater entrepreneurial surplus.Market participants would reason that because the issuer’s entrepreneurschoose to reveal more, the issuer probably has better prospects relative toissuers disclosing less.

Thus it is the fact that an issuer chooses to disclose, not disclosureitself, that leads to the association between greater disclosure and highershare price. This concept is the basis of signaling theory: issuers that have goodnews signal this fact by disclosing their news and those that do not have goodnews signal this fact by their inability to make comparable disclosures.42

While the signaling phenomenon means that the market will be betterinformed in a system of issuer choice than might first appear, it will not be as wellinformed as if all issuers are compelled to disclose at the higher level that someissuers choose voluntarily. Silence is not a complete substitute for affirmativelydisclosing lack of good news because the market knows that there are reasonswhy an issuer would choose a low disclosure regime besides lack of goodnews.43 As we have seen, an issuer may choose not to disclose because


able, from the silence of the issuers involved, even to begin to infer in advance that NewYork City and the Washington Public Power System would experience disastrous defaults.Id. These were the two largest defaults in the history of the United States, but the issuers,as municipal entities, were exempt from the mandatory disclosure system. See Ann J. Gellis,Mandatory Disclosure for Municipal Securities: A Reevaluation, 36 BUFFALO L.REV. 15(1987). 44. A second reason, other than lack of good news, why an issuer might not discloseis that management is contemplating a management buyout and does not want the goodnews it possesses to increase the share price. This possibility is particularly damaging tothe signaling mechanism since it suggests that silence might in fact indicate highly positivenews. See Coffee, supra note 7, at 740-741. This problem is likely to be relevant, however,only where we are looking to signaling to produce price accuracy in the secondary market,not in the IPO market. It also relates to behavior that is partially constrained by rulesagainst insider trading. 45. Romano, supra note 1, at 2374.

revealing the information might put it in an inferior position vis a vis a competitor,major supplier or major customer.44

Projects therefore are not as well chosen in a signaling world–whereissuers, through choice of regime, are free to choose the level at which they bindthemselves to disclose and the market draws negative inferences from thedecision of the issuers that choose a low level regime–as they would be chosenin a world where all issuers disclose at a high level. Moving from the first worldto the second produces social benefits because the list of projects implementedwould be closer to the ideal. If these social benefits exceed the social costs, themove represents a net social gain. The entrepreneurs that do not disclose at ahigh level in the first world would not, through higher prices, fully capture thisgain from their increased disclosure in the second world. Part of the gain wouldinstead accrue to the entrepreneurs who disclose at a high level in both worldsbecause the improved allocation of capital would mean a higher percentage oftheir projects would be implemented. In essence, to price each IPO properly,the market needs information about all potential projects so that it can make therelevant comparisons accurately.

Professor Romano cites a number of studies showing a positiverelationship between disclosure and share price as evidence that there are“powerful incentives” for issuers seeking new funds to disclose information andthat issuers can hence be trusted to choose for themselves the socially optimaldisclosure regime.45 This relationship, we have just seen, is due to the fact thatthe issuer chooses to disclose, not the disclosure itself. It is the pattern we wouldexpect to find under signaling theory. In an issuer choice world where we relyon signaling, there will be incentives for an issuer to choose a regime requiring alevel of disclosure greater than zero. But, as we have just seen, these incentives


46. See Easterbrook & Fischel, supra note 7, at 685. The model by Ronald Dye, cited byRomano in her effort to dismiss the divergence between private and social costs causedby interfirm costs, is relevant here as well. Dye, supra note 17. Dye refers to disclosurethat is useful for analyzing another issuer as involving a “financial externality.” While Dyesuggests that where only financial, not real, externalities are involved, voluntary disclosurewould be socially optimal under a fairly wide range of circumstances, id at 15, a number offactors render the model not a useful guide to policy. See supra note 18. 47. Under the current regime, post-IPO disclosure comes from two sources. One sourceis the periodic disclosure requirements under Sections 12, 13 and 15(d) of the ExchangeAct. This disclosure is provided on Form 10-K (an annual report requesting much of thesame information as is requested on the Securities Act Form S-1 used in connection withIPOs), Form 10-Q (a quarterly report) and Form 8-K. The other source is the Securities Actrequirements to register each post-IPO public offering of securities. Because the issuer andits directors and officers are much more likely to face substantial liability where there is amaterially false or misleading statement or omission in a Securities Act registrationstatement than in an Exchange Act filing, and because under the Securities Actunderwriters face strict liability for such statements or omissions unless they candemonstrate that they engaged in due diligence, issuers provide higher quality disclosurewhen they engage in subsequent public offerings than when they do not. See Merritt B.Fox, Shelf Registration, Integrated Disclosure, and Underwriter Due Diligence: AnEconomic Analysis, 70 VA. L. REV. 1005, 1025-30 (1983) (hereinafer Shelf Registration);Merritt B. Fox, Rethinking Disclosure Liability in the Modern Era, 75 WASH. U. L. Q. 903,903-4 (1997) (hereinafter Rethinking Disclosure Liability). Because, as analyzed below,greater disclosure more effectively constrains managers from engaging in non-share-maximizing behavior, firm managers that anticipate making such offerings tend to act morein concert with the interests of shareholders. See Frank Easterbrook, Two Agency CostExplanations for Dividends, 74 AM. ECON. REV. 650 (1984); MERRITT B. FOX, FINANCE AND

INDUSTRIAL PERFORMANCE IN A DYNAMIC ECONOMY: THEORY, PRACTICE, AND POLICY 121-140(1987) (Hereinafter FINANCE AND INDUSTRIAL PERFORMANCE).

will not be great enough to induce issuers to choose a regime requiring a level ashigh as is socially optimal, even in the IPO context.

This “public goods” aspect of issuer disclosure does not end here. Thereare other ways in which information disclosed by one issuer about itself can beuseful in analyzing other issuers. It could, for example, reveal something aboutpossible industry wide trends.46 Again, these are social benefits that the issuerdisclosing the information cannot appropriate through higher share price.

3. Greater Managerial Adherence to Shareholder Interests.

A third benefit of issuer disclosure is a reduction in the extent to whichmanagers of public corporations place their own interests above those of theirshareholders. Here I am talking primarily not about disclosure at the time of anIPO, but about what is provided thereafter.47 Greater ongoing, periodicdisclosure increases the effectiveness of a number of devices that work to limit


48. The reduction in managerial discretion is, as discussed above, a direct cost tomanagers, but this cost is smaller than the social gains resulting from better resourceallocation. Michael Jensen & William Meckling, Theory of the Firm: Managerial Behavior,Agency Cost, and Ownership Structure, 3 J. FIN. ECON. 305 (1976). The question here is theextent to which managers feel these social gains. 49. I discuss these points in more detail in Merritt B. Fox, Required Disclosure andCorporate Governance, in KLAUS I. HOPT, ET AL., EDS., COMPARATIVE CORPORATE

GOVERNANCE: THE STATE OF THE ART AND EMERGING RESEARCH 701-718 (1998). In theUnited States, we are so accustomed to a high level of issuer disclosure that we tend notto appreciate its importance with respect to these devices. A comparison with Russia isrevealing. The dearth of disclosure there renders the fiduciary duties nominally imposedon management almost useless. See Bernard Black & Reinier Kraakman, A Self-EnforcingModel of Corporate Law, 109 HARV. L. REV. 1911 (1996). It also makes relativelymeaningless disinterested shareholder approval of transactions in which management isinterested. See Merritt B. Fox and Michael A. Heller, Corporate Governance Lessons fromRussian Fiascos; (unpublished paper presented at the American Law and EconomicsAssociation Annual Meeting, May 1998, on file with the author). 50. The market for corporate control is a well-recognized device for limiting the agencycosts of management where ownership is separated from control, as in the typical publiclyheld corporation. More information and the resulting increase in price accuracy improvesthe control market’s effectiveness in performing this role. A potential acquirer, in decidingwhether it is worth paying what it would need to pay to acquire a target that the acquirerfeels is mismanaged, must make an assessment of what the target would be worth in theacquirer’s hands. This assessment is inherently risky and acquirer management is likelyto be risk averse. Greater disclosure, however, reduces the riskiness of this assessment.Hence, with greater disclosure, a smaller apparent deviation between incumbentmanagement decisionmaking and what would maximize share value is needed to impel apotential acquirer into action.

Also, when share price is inaccurately high, even a potential acquirer that believes forsure that it can run the target better than can incumbent management may find the targetnot worth paying for. The increase in share price accuracy that results from greaterdisclosure reduces the chance that a socially worthwhile takeover will be thwarted in thisfashion.

Greater disclosure thus makes the hostile takeover threat more real. Incumbent managerswill be less tempted to implement negative net present value projects in order to maintainor enlarge their empires, or to operate existing projects in ways that sacrifice profits tosatisfy their personal aims. Those that nevertheless do these things are more likely to bereplaced. See, FOX, FINANCE AND INDUSTRIAL PERFORMANCE, supra note 47, at 84-91.

such behavior.48 Disclosure assists in the effective exercise of the shareholderfranchise and in shareholder enforcement of management’s fiduciary duties.49

Even more important, disclosure increases the threat of hostile takeover whenmanagers engage in non-share-value-maximizing behavior. This is becausedisclosure both makes a takeover less risky for potential acquirers and reducesthe chance that a value-enhancing acquisition will be deterred by the targethaving an inaccurately high share price.50


51. See supra note 46 and accompanying text. 52. See Easterbrook & Fischel, supra note 7, at 684. 53. Professor Romano suggests as a default rule that management would not be allowedto change disclosure regimes without the approval of a majority of shareholders. Romano,supra note 1, at 2415-16. She also suggests that any individual issuer could choose torequire a higher, supermajority vote. Id. While these suggestions might appear to answermy concern, in reality, they pose a dilemma. On the one hand, relying on shareholdervoting to protect against managerial opportunism involves the classic free rider problem.For a shareholder with a small fraction of the shares of a corporation – the type that inaggregate hold the majority of the shares of most corporations – it is not worth acquiringthe information to judge the desirability of such a change in regime. This is because thereis only a very small chance that her vote will affect the outcome of the elections and evenif it did she would experience only a small fraction of the effect of the change in regime. SeeBernard S. Black, Shareholder Passivity Re-examined, 89 MICH. L. REV. 520 (1990); JeffreyN. Gordon, The Mandatory Structure of Corporate Law, 89 COLUM. L. REV. 1549, 1573-75(1989). On the other hand, if the percentage needed for approval is set very high to help

In determining the extent to which an issuer’s entrepreneurs andmanagers will be able to capture the social benefits from its ongoing, post-IPOdisclosure, we should start by observing that, like IPO disclosure, thisinformation too has “public goods” aspects.51 It will aid in the analysis of theprospects of other issuers just going public with IPOs. It will also help disciplinemanagers of other established public issuers by assisting the devices that limit theability of their managers to deviate from their shareholders’ best interests. If onehas detailed information about one issuer’s performance, for example, it is easierto detect shirking by the managers of its competitors, who face a similar externalbusiness environment. These public goods type benefits will not be captured inthe price of the issuer making the disclosure.

As for the disciplining effect of any given level of periodic disclosure onan issuer’s own management, there is, in theory, a way for its managers tocapture the full social benefit. They do this by binding the issuer at the time ofits IPO to provide in the future periodic disclosure at that level. The higher thepromised level, the less the market would expect management decisionmakingto deviate from what is in shareholders’ best interests, and the higher the marketprice for the issuer’s initial offering, net of the prospective costs of thiscommitment. The entrepreneurs capture this price improvement through adollar-for-dollar increase in the size of the entrepreneurial surplus.52

Reality will be different. This internalization of the benefits from thedisciplinary effects of disclosure will not be complete unless the market isconfident at the time of the IPO that the issuer is totally bound to disclose at thechosen level for the life of the firm. It is probably impractical to institute an issuerchoice regime with such an ironclad guarantee.53 Yet


protect against this problem, the arrangement locks the issuer in with a disclosure regimethat is not likely to adapt to the particular changing needs of firms from the issuer’s homecountry. This is in contrast to a system of mandatory disclosure based on the issuer’snationality, where the issuer’s disclosure would be governed by a regime that is likely tomake these adaptations.

Romano also argues that the cost to an institutional investor of informing itself for anygiven vote is lower than it might first appear. This is because the question of the effect ofa switch to any particular regime or share value is likely to arise with respect to many of theissuers in which the institution holds shares. This is a fair point. It is limited, however, tothe extent that arguments in favor of such a switch differ in their validity from one issuerto the next and do so in ways the determination of which requires information about theparticular issuers involved.

There is also an open question as to how representative the preferences of institutionalinvestors are of the preferences of other shareholders on the question of how much anissuer should be required to disclose. On the one hand, it seems clear that institutionshave greater skill and inclination than do small individual investors at obtaining informationabout an issuer from sources other than the issuer’s public disclosures. Romano arguesthat studies showing that institutional investors do not outperform the market suggestotherwise, but this is not correct. Such studies show only that the market reflects newpublicly available information quickly so that it is hard to make trading profits frompossessing it. They do not show that investors of all kinds possess the same amounts ofinformation. If an issuer fails to provide high disclosure the, greater skill and inclinationof institutional investors at obtaining information means that their cost of effectivelyexercising their franchise is smaller than that of the individual investor. Institutions canmore easily find substitutes to issuer disclosure for obtaining information and so theirpreference for a high level of issuer disclosure would not be as intense. In fairness, itwould be easy to exaggerate the importance of this difference, however. The collectiveaction problems for small individual investors may be so great that disclosure is of nodirect use to them anyway, since they would not make the effort to process the informationeven if it was given to them. Disclosure benefits them indirectly instead by making moreeffective the other constraints on management.

Professors Choi and Guzman propose that an issuer select a securities law regime at thetime of each new issue of securities. Choi & Guzman, Portable Reciprocity, supra note 3,at 922. Since they propose no exit mechanism, the issuer appears to be bound by thatchoice – including its ongoing periodic disclosure requirements – for as long as thesecurities are outstanding. This approach solves the commitment problem in an ironcladway, but it has, in an even more extreme form, the adaptability problems that a highsupermajority clause would have under Romano’s proposal. See infra Part IV.

without it, the market knows that managers will be subsequently tempted toswitch to a lower disclosure regime and will discount the share price at the timeof the IPO to reflect this possibility.

This temptation arises because lower disclosure reduces theeffectiveness of the devices that limit managerial discretion and hence providesmanagers with room to make more decisions that satisfy their own objectivefunctions at the expense of the interests of shareholders. Managers who


54. In other words, the switch is partially paid for in advance, whether or not it is made.As a result, the cost to management of actually making the decision to switch is lowered,thereby making the switch more probable. This is a classic example of a “lemons” problem.George Akerlof, The Market for “Lemons”: Quality Uncertainty and the MarketMechanism, 84 Q.J.ECON. 488 (1970). 55. Managerial shareownership and stock options can ameliorate, but not eliminate thisproblem. Such holdings constitute only a fraction of the issuer’s outstanding shares–inmost cases a small fraction–and so most of the reduction in share value from non-share-value maximizing decisions is externalized onto other persons. 56. As a result, the case, based on signaling theory, that issuer choice will lead tosocially optimal disclosure is inherently even harder to make for periodic disclosure thanfor IPO disclosure. Steven Ross, in his classic exposition of the theory, assumes thatmanagers are a dollar better off for every dollar by which they increase the value of the firm.Ross, supra note 6, at 185. Telling a convincing story as to why this might be so in thecase of periodic disclosure is much more complicated than in the case of IPO disclosure,since, unlike in an IPO, the entrepreneurs are not in essence selling a portion of the equity

consider switching to a lower disclosure regime know that if they do, their shareprice in the secondary market will decline because of the expected reduction inmanagerial discipline. The decline will not necessarily be as much, however, asthe reduced managerial discipline diminishes the share value. This is because, asjust noted, the market knows in advance that managers will be tempted to switchto a lower disclosure regime and the discount reflecting this possibility will still bepresent at the time that any actual switch occurs. In such a situation, the pricedrop if a switch did occur would be less than the drop in share value. Theswitch would therefore be less costly to managers and thus more likely.54

Moreover, even if the price decline were as great as the decline in value,the managers would often find that the increased freedom, especially thatresulting from the reduced threat of takeovers, would be worth the pricedecline.55 After all, their biggest concern with secondary market share price isoften the takeover threat in the first place. Admittedly, the switch would alsoincrease the issuer’s cost of seeking additional capital through new share issues.Many established firms, however, never raise new capital in this fashion. Evenfor those that do, the cost of the lower share price will largely or wholly be borneby the existing public shareholders, not the managers.

In sum, the lack of an iron clad guarantee combined with this temptationmeans that when managers, at the time of the IPO, choose the regime underwhich their issuer will operate, they are not likely to capture fully the beneficialdisciplining effects of their promised level of disclosure. And if they subsequentlychoose a less rigorous regime, the suffering they incur as a result of any pricedrop at the time of the switch is not likely to be as great as the decline in sharevalue due to reduced discipline.56


of the company previously belonging to them. Professor Coffee, for example, observes The theory of voluntary disclosure does seem to have some validity as applied to initialpublic offerings and, to a lesser extent, to all primary distributions. This theory has farless persuasive force, however, when applied to secondary market trading, which the ‘34Act chiefly governs. Here high agency costs currently exist . . . , thus shelteringopportunistic managerial behavior. A management that opposes a lucrative takeoveroffer to its shareholders is also capable of biasing its disclosures. (footnotes omitted)

Coffee, supra note 7, at 746-47. 57. In theory an issuer whose economic center of gravity as a firm is in the United Statescould avoid the U.S. disclosure regime if it offered its shares only abroad, listed its sharesfor trading only abroad, incorporated abroad and had a majority of its shares held byforeigners. Few, if any, such issuers exist, however. See Fox, Securities Disclosure in aGlobalizing Market, supra note 14, at 2608-2618.

II. EMPIRICAL EVIDENCE CONCERNING THE WELFARE EFFECTS OF

MANDATORY DISCLOSURE

Under the current system, managers of U.S. issuers are required todisclose at the level prescribed by the U.S. authorities.57 Under issuer choice,they would be free to bind themselves to the disclosure level of their choosing.The immediately preceding Part demonstrates that a system of issuer choicewould involve substantial market failures. Thus a switch to issuer choice willdecrease social welfare unless the current system of mandatory disclosureinvolves governmental failures that are even larger. This observation gives riseto two questions. First, what empirical evidence is available concerning therelative impact on social welfare of the current system versus issuer choice?Second, what are the implications of this evidence for whether the United Statesshould switch?

A. Empirical Evidence Concerning the Imposition of MandatoryDisclosure.

There are no studies comparing the current system with issuer choice,since issuer choice has never been tried. There are studies, however, comparingthe current system with the previous world of no federal regulation.

1. Studies of the Effect of New Issue and Periodic Mandatory Disclosureon Price Accuracy.

The efficient market hypothesis suggests that an issuer’s share price willbe an unbiased estimate of the share’s actual value whether there is a lot of


58. See supra I.C.2. 59. See supra I.C.2. To put this concept of expected accuracy in statistical terms,consider price to be a random variable generated by a distribution function with a meanequal to actual value (reflecting the fact that the price is unbiased). A good measure of theprice’s expected accuracy would then be the variance of the distribution – the expectedvalue of the square of the deviation from actual value. The greater the variance, the lowerthe price’s expected accuracy. 60. George Stigler, The Economics of Information, 69 J. POL. ECON. 213 (1961). 61. George Stigler, Public Regulation of the Securities Markets, 37 J. BUS. 117, 122, 123tbl.3. (1964) (hereinafter Stigler, Public Regulation). 62. This assumes, in accordance with the rational expectations model, that the market fornew issues, whether it is provided with a lot of disclosure or only a little, is efficient in thesense that each issuer’s price is an unbiased estimate of its actual value. See supra note40 and accompanying text, and infra note 89.

publicly available information about the issuer or only a little.58 Moreinformation, however, will increase the expected accuracy of the price, i.e., thelikelihood that the price is relatively close, one way or the other, to actualvalue.59 This is because with more information, speculators–the persons who setan issuer’s share price in the market–have a more accurate sense of the issuer’sfuture. Accuracy is revealed empirically by the amount of dispersion in theissuer’s share price over time. To quote George Stigler, “[p]rice dispersion isa manifestation–and, indeed, it is the measure–of ignorance in the market.”60

Thus we would expect that if the imposition of the federal mandatory disclosuresystem led to an increase in meaningful information being publicly available, itwould have reduced price dispersion. There is considerable evidence that pricedispersion did in fact decline after imposition of the federal regime.

a. Studies by Stigler and others of the effect of new issuermandatory disclosure on price accuracy. George Stigler examines two groupsof new share issues, one from the period 1923-28 (prior to the passage of theSecurities Act of 1933, with its new issue disclosure requirements) and the otherfrom the period 1949-55 (after the Act’s passage). For each group, hecalculates the average of the price performance of the group members over thefive years after issuance relative to the price performance of the market as awhole for the corresponding period (thus controlling for factors affecting themarket as a whole). He finds that the variance in the relative price performanceof individual share issues around the average of the group as a whole declinedby almost half between the pre-Act group and the post-Act group.61 This resultis thus consistent with the proposition that the improved disclosure associatedwith new issuers after adoption of the ‘33 Act led to greater price accuracy62

since, compared to the pre-Act group, the initial price of the typical issuer in thepost-Act group (discounted to present value and adjusted to account for the


63. Romano sites as her first explanation of this finding that “the legislation simplyforced riskier investments off the market.” Romano, supra note 1, at 2377 (footnotesomitted). Romano admits, however, that “this finding can also be “interpreted as indicatingthat the disclosure mandated by the Act enabled investors to form more accurate pricepredictions.” Id. Her first explanation is an idea originally put forward by Stigler. Stigler,supra note 61, at 122.

Romano and Benston’s first explanation is not the explanation adopted by others whohave looked at the data. See Irwin Friend & Edward S. Herman, The S.E.C. Through aGlass Darkly, 37 J. BUS. 382, 390-91 (1965); Carol J. Simon, The Effect of the 1933 SecuritiesAct on Investor Information and the Performance of New Issues, 79 AM. ECON. REV. 295,311-31 (1989). Presumably everyone, including Stigler and Romano, accepts the theoreticalproposition that any information that is of value to investors for predicting the future withgreater accuracy will lead to less share price dispersion. The results showing that the post-Act group in fact had less dispersion should therefore logically lead one to the conclusionthat the information which the Securities Act prompted to be disclosed was in fact of suchvalue, unless one has affirmative evidence that leads one to believe that some other factorwas responsible. The only affirmative evidence offered by Romano is a citation to a studyshowing an increase in the proportion of financing by debt after passage of the Act,particularly in the form of privately placed, higher risk debt. Romano, supra note 1, at 2377.This development would not seem surprising in the depths of the Great Depressionregardless of whether mandatory disclosure was imposed on new issues or not.

Moreover, there is no obvious reason why increased disclosure, which simply reducesinformation asymmetries between managers and the market, would in fact force riskierinvestments off the market. It would hurt the chances of riskier projects that, based on allavailable information including what is known by the managers proposing each project,should not be implemented. But it would help the chances of riskier projects that, basedon the same information, should be implemented. See supra I.C.2 for a discussion aboutthe relationship between disclosure and which new projects go forward.

It may be possible to tell a story of how the liability system that helps enforce theSecurities Act disclosure requirements is biased against riskier projects. Presumably, thestory would suggest that potential issuers with riskier projects and persons associatedwith such issuers face a greater risk of liability, even if they make just as great an effort tocomply with the disclosure requirements, than potential issuers with less risky projects andtheir associated persons. Perhaps to suggest such a story, Romano cites an article byTinic, which argues that underwriters, fearing liability, switched to larger, less risky issuersafter adoption of the Act. Shea M. Tinic, Anatomy of Initial Public Offerings of CommonStock , 48 J. FIN. 789 (1988). Tinic uses indirect evidence that this is what happened: afinding that the offering price discount relative to initial trading prices was higher in a post-Act 1966-71 sample of IPOs than in a pre-Act 1923-30 sample. Id. at 805. The largerdiscount reduces both the probability of litigation and, if it occurs, the amount of damages.He theorizes that the discount therefore constitutes a form of litigation insurance, id. at797-803. Thus Tinic’s theory that there has been a switch to less risky issuers rests on theclaim that the increase in the discount represents litigation insurance. A number of factorsundermine the force of this explanation of the increased discount and hence its probity asevidence that there was a switch to less risky issuers because of fear of litigation. First, the

performance of issuers generally) was closer to what the issuer’s performancei n f a c t t u r n e d o u t t o b e . 6 3 T h e


cost of this insurance would be extraordinary – in essence one gives up a dollar today inorder to be sure that one is not sued for that same dollar later – and so it is hard to believethat rational contracting parties would agree to such an arrangement. Second, there aresevere problems in attributing the increase in the discount to this one change between thetwo sample periods when so many other things have changed as well. Finally, there areerrors in Tinic’s description of the potential litigation costs created by the Act thatdiminish the force of his litigation insurance theory as an explanation of the discount. SeeJanet Cooper Alexander, The Lawsuit Avoidance Theory of Why Initial Public Offeringsare Underpriced, 41 U.C.L.A. L. REV. 17, 26 n. 28 (1993).

Beyond Romano’s citation of Tinic, neither Stigler nor Romano have attempted to tella story to explain why the Security Act’s liability system is biased against riskier projects,or to suggest why the bias that such a story would predict is so great as to be a moreplausible explanation for the observed reduced price dispersion than that the informationdisclosed helped investors better predict the issuer’s future. 64. Simon, supra note 63, at 306-306 & tbl.4 (1989). Use of several post-Act periods rulesout the “bull market” explanation of Stigler’s results, i.e., that the greater variance in thepre-Act period was due to the fact that it was a boom period whereas the post-1933 periodthat Stigler used was not. Id. at 308-313. 65. George Benston, Required Disclosure and the Stock Market: An Evaluation of theSecurities Exchange Act of 1934, 63 AM. ECON. REV. 132 (1973). The residual is thedifference between the issuer’s actual return in a month (price change plus dividends) andthe return predicted by the market model (the market model predicts an issuer’s returnbased on overall market’s performance during the month and the issuer’s traditionalsensitivity to market-wide trends). Thus the variance of an issuer’s residuals is a measureof how much uncertainty exists about an issuer relating to factors specifically affecting theissuer, i.e., it is a measure of firm-specific or non-systematic risk. A reduction in thevariance of a firm’s residuals would suggest an increase in meaningful information in themarket about the firm.

second, more recent study, by Carole Simon, makes a similar kind ofcomparison, using the techniques of modern financial economics. This time thecomparison is between a group of pre-Act new share issues and groups of newshare issues in each of several post-Act periods. Like Stigler, she finds a lowervariance for the share issues made after mandatory disclosure was imposed thanfor the earlier ones.64

b. Benston’s study of the effect of periodic mandatory disclosure onprice accuracy. George Benston studied the effect of the periodic disclosurerequirements under the Exchange Act. His results are more mixed than the newissue disclosure studies. Benston looked at 466 New York Stock Exchange(NYSE) firms for a period starting prior to the imposition of the Exchange Actregime and running to a point 90 months after its imposition, using the marketmodel to determine the variance of their month to month residuals as a measureof price dispersion.65 Of these, 290 firms disclosed sales data before impositionof the Exchange Act regime and 176 did not. After imposition, all disclosedsales data. Benston focused on whether the riskiness of the 176 firms that were


66. Id. at 148-49 and tbl. 4. This, of course, does not prove that there was no suchdecline. This part of Benston’s study looks at a single disclosure item among a wholepackage of new requirements. See infra notes 70-79 and accompanying text. The effect ofthis single requirement on price accuracy may have been positive and large enough relativeto its costs to make it economically worthwhile, but not large enough, given the relativelyweak statistical powers of the test, to show up as significant. See infra II.A.3.

There are also some problems with the fundamental design of Benston’s test. First, the“control” group – the firms that were already reporting sales – themselves significantlyimproved the quality of their sales disclosures in response to the regulations. Prior to thepassage of the Exchange Act, there was great variation among these firms in how the salesfigure was calculated. See Maurice C. Kaplan & Daniel M. Reaugh, Accounting, Reportsto Stockholders, and the SEC, 48 YALE L.J. 935, 948 (1939), and infra notes 76-79 andaccompanying text. This improvement in the quality of the sales disclosures of thedisclosing firms would mute any differences in the price dispersion reactions of the twogroups to imposition of the sales disclosure requirement even if the requirement diddecrease the dispersion of the previously non-disclosing firms. It would therefore be lesslikely that any difference in reaction would show up in the data and be statisticallysignificant.

Second, a test looking at the reaction in the 1930s of share price riskiness to theimposition of Exchange Act periodic disclosure requirements is less likely to show thepositive effects of mandatory disclosure than a test, such as those discussed in the textabove, looking at the reaction of share price riskiness to imposition of the Securities Actnew issue disclosure requirements. This is because issuers had fewer incentives in the1930s to provide accurate information in response to the Exchange Act’s requirements thanthey had in response to the Securities Act’s requirements. The Securities Act had from thevery beginning extensive civil liability provisions under Section 11 (imposing absoluteliability on the issuer and, unless they would affirmatively prove they performed adequatedue diligence, on the directors, key officers and underwriters as well) and under Section12(a)(2). There was also regular SEC staff review of S-1 filings. The Exchange Act had onlySection 18 liability, whose insuperable reliance requirements have made it a dead letter. SeeDAVID L. RATNER & THOMAS LEE HAZEN, SECURITIES REGULATION CASES AND MATERIALS

330-331 (5th ed. 1996). There has been no regular agency review of Exchange Act filings.The incentives to comply with Exchange Act periodic disclosure requirements are much

greater today than then. Thus there is a particular danger in inferring that today the Acthas no impact on price dispersion from a finding that there was no statistically significantextra reduction in price dispersion among the nondisclosing firms back in the 1930s whenthe Exchange Act’s sales disclosure requirements were imposed. The increase in incentives

required by the Act to reveal their sales figures for the first time (the “non-disclosing firms”) declined relative to riskiness of the firms that had beendisclosing sales all along (the “disclosing” firms). To do this, he looked to seewhether, after the Act was imposed, the average decline in the variance of theresiduals of the non-disclosing firms was greater than that of the disclosing firms.He found that it was, but only by a very small, statistically insignificant amount.He concluded from this that the post-Act disclosure of sales data by the non-disclosing firms did not reduce their riskiness.66


today is due in part because of developments under Exchange Act Section 10(b) and Rule10b-5. Potential Rule 10b-5 liability for Exchange Act disclosure violations by firms that donot trade in their own securities did not develop until the late 1960s, S.E.C. v Texas GulfSulphur, 401 F.2d 833 (2d Cir. 1968). This potential liability in turn did not become a seriousthreat to most issuers until class actions became possible with the development of thefraud on the market theory of reliance, which was first enunciated in the lower courts in the1970s and was affirmed by the Supreme Court only in 1988. Basic Inc. v Levinson, 485 U.S.224 (1988). The advent of integrated disclosure in the early 1980s, in which many issuerscan incorporate by reference their Exchange Act periodic filings into their new issueSecurities Act registration statements, also created an incentive for improving the qualityof Exchange Act disclosure not present in the 1930s. This is because the itemsincorporated by reference become subject to the Securities Act liability scheme. I explorethe shrinking difference in incentives between the Securities Act and the Exchange Actdisclosure regimes in greater detail in Fox, Shelf Registration, supra note 47 and in Fox,Rethinking Disclosure Liability, supra note 47. 67. Benston, supra note 65, at 148-49 and tbl. 4. See also Irwin Friend & RandolphWesterfield, Required Disclosure and the Stock Market: Comment 65 AM. ECON. REV. 467(1975). 68. Benston, supra note 65, at 142. 69. Romano, supra note 1, at 2373.

Benston’s results are also relevant, however, as to the value of the totalpackage of information required to be disclosed under the 1934 Act regime: theriskiness of each group, as measured by the average standard deviation of themembers’ residuals, declined by about one third from the pre-Act period to thepost-Act period.67 This would imply that whatever the effect of just the newlymandated sales data, the total mandated package did substantially increase thesupply of meaningful information in the market and, as a consequence, improveprice accuracy. Benston, however, ignores these results of his own study.

Why do Romano and Benston ignore the fact that each group’sriskiness–the non-disclosing firms and the disclosing firms–appears to havedropped substantially after the imposition of the Act and focus exclusively on thedifference between the amount by which each dropped? The reason is that theyconclude in advance that the Act’s other disclosure requirements were of noimportance, even in the aggregate. Benston describes his study as one that teststhe effect of the Act “by examining its differential effect on the securities ofcorporations that were and were not affected by the legislation [the non-disclosing firms versus the disclosing firms].”68 Romano cites Benston for theproposition that the “only major mandated item that was not reported by asignificant set of firms prior to the 1934 legislation was sales.”69

Benston and Romano’s conclusion that the rest of the Exchange Act’sdisclosure requirements were of no importance would have come as a realsurprise to commentators at the time of the Act’s passage. To start, it rests on


70. Benston, supra note 65, at 142. 71. The standard agreement, at least for issuers that applied in the mid 1920s or later,obligated the issuer to provide certain kinds of information on an ongoing basis. See JohnHanna, The Securities Exchange Act as Supplementary of the Securities Act, 4 LAW AND

CONTEMPORARY PROBLEMS 256, 258-259 (1937). See infra note 73 for the practice prior tothat time. 72. Id. at 259. 73. The New York Stock Exchange’s Rules of Listing dated July 1, 1925 (which, sincethey were included in a 1930 casebook of an eminent scholar, were apparently still in forcein 1930) is reproduced in ADOLPH A. BERLE, JR., CASES AND MATERIALS IN THE LAW OF

CORPORATION FINANCE 700-713 (1930). The same rules dated September 20, 1938 arereproduced in in W ILLIAM H. BLACK, THE LAW OF STOCK EXCHANGES, STOCKBROKERS AND

CUSTOMERS 997-1027 (1940). The listing agreement required under the earlier dated rulesdid not require that the financial statements be audited nor that non-recurrent items ofincome be disclosed as such.

The earlier that an issuer in Benston’s study listed on the NYSE, the less strict were theperiodic disclosure requirements under which it was operating during the pre-Act testperiod. The Twentieth Century Fund, in its exhaustive 1935 study of the securities markets,concluded:

It was, however, not until the boom period ending in 1929 was well under way that theExchange began in earnest to develop the requirements for disclosures that it todayimposes upon corporations desiring to list their securities. Since the depression ... theExchange authorities have given further attention to this matter.

TWENTIETH CENTURY FUND, INC., THE SECURITIES MARKETS 577 (1935). Similarly, ProfessorHana stated in 1937 that “listing requirements have become progressively more rigorousin recent years.” Hana, supra note 71, at 258. There is some reason to believe that issuersthat listed prior to the mid 1920s had no obligation at all to provide ongoing periodicinformation. See TWENTIETH CENTURY FUND, INC., STOCK MARKET CONTROL 133 (1934).

the faulty premise that everything else required by the Exchange Act was alreadyrequired by the exchanges. Benston’s method of researching what theexchanges required in the 1920's and early 1930's was to make inquiries ofthem. He was apparently informed that during this period the exchanges already“had . . . rules that required listed companies to send certified income statementsand balance sheets to stockholders in advance of the annual meeting.”70 In fact,there were no such rules. What regulation there was came from the fact thateach issuer, at the time it applied to have its securities listed on the New YorkStock Exchange (NYSE), was required to enter into a listing agreement with theExchange.71 It is true that about the time of the passage of the Exchange Act,the NYSE started to require that each new applicant enter into a listingagreement that provided that a certified income statement and balance sheet begiven annually to its shareholders.72 But the listing agreements required of issuersthat applied prior to that point (which would include all the firms in Benston’sstudy) did not contain requirements this strict, in some cases not nearly thisstrict.73 Indeed, one of the problems addressed by the Act was the fact that the


To show how little bound many firms that listed prior to 1925 apparently were (or,alternatively, how lax was the enforcement of the listing agreements generally), considerthe results of a survey conducted by Kaplan and Reaugh. They made a careful examinationof 70 firms’ annual financial reports to shareholders for 1930 (a year in the middle Benston’spre-Act test period). Kaplan and Reaugh characterize these 70 firms as representative ofthe 500-600 largest non-utility, no-railroad, non-financial corporations in America. Kaplan& Reaugh, supra note 66, at 938 n. 16. Four of these seventy did not provide theirshareholders with even the simplest form of income statement. Id. at 940. A majority ofthe surveyed firms did not report “cost of goods sold.” Id. at 948. 74. S. REP. NO. 792, 73rd Cong., 2d Sess. 5 (1934). See also John E. Tracy & Alfred B.MacChesney, THE SECURITIES EXCHANGE ACT OF 1934, 32 MICH. L. REV. 1025, 1049 (1934)(the “New York Stock Exchange, through its listing requirements, has made commendableprogress in setting higher standards” but that these measures do not “reach securitiesalready issued and listed.”) Professors Tracy and MacChesney describe the disclosurerequirements of the Exchange Act as “fill[ing] in the gap left open by ... the listingrequirements of the exchanges; they secure similar information [to that required by theSecurities Act] regarding securities already outstanding.” Id. at 1053. 75. For example, the Exchange Act required information that “differ[s] from the usualstock exchange listing requirements in [its] demand for disclosure of the stock interests andremuneration of officers, directors, ten percent stockholders and others. Hana, supra note71, at 261. Also, almost one third of New York Stock Exchange companies publishedfinancial statements only once a year. TWENTIETH CENTURY FUND, INC., STOCK MARKET

CONTROL 135 (1934). The SEC required issuers to provide quarterly financials, pursuant toForm 10Q under the Exchange Act. 76. See Kaplan & Reaugh, supra note 65, at 940 (describing how the 1930 annual incomestatements of 34 of the 70 large corporations surveyed “were obviously inadequate” andhow, as a group, the income statements of all the surveyed corporations displayed a“striking lack of uniformity as to form and content.”) 77. See John Hana, The Securities Exchange Act of 1934, 23 CALIF. L. REV. 1, 20 (1934).The SEC, in describing its Exchange Act disclosure requirements, commented specificallyon the shortcomings of listing agreement disclosure:

The outstanding advances which these requirements [under the Exchange Act]represent over reporting practices already in vogue are first, a greater emphasis on theaccounting steps involved in income determination and second, a more completeexplanation of the changes which have occurred in balance sheet items during the yearunder report. .. The requirements give less attention, on the other hand, to historical

exchanges, according to a Senate report, felt hampered in their ability to inducesuch earlier listing firms to provide more information because they “did notbelieve themselves entitled to modify [these agreements] without the consent ofsuch issuers.”74

Certain important items of information were not required at all under thelisting agreements, even those of firms listing just before passage of the Act.75

More importantly, there were enormous problems with the of the quality of thefinancial data that was being provided under the listing agreements.76 A primarypurpose of the Exchange Act was to remedy these problems.77 One example


information concerning the company, since all companies affected are already listed onthe exchanges, and have been reporting consistently, under existing exchangerequirements.

Securities Exchange Act Release No. 66 (December 21, 1934). 78. Kaplan and Reaugh, supra note 66 at 938 n.16. 79. Id. at 956. 80. Romano, supra note 1 at 2773. 81. A second reason for Benston’s focus on sales was that he did a separate studypurporting to show that in 1964, among different types of accounting numbers reported inSEC filings, “sales [were] the only relatively important accounting number.” Benston,supra note 65, at 138-39. He measured the importance of an accounting number by whetheran unexpected change in the number from the number previously disclosed led tosignificant share price reaction. Id. at 138-39. A finding that only the sales numberqualifies as important by this measure hardly suggests that there is no value in requiringall the other information required by the Exchange Act. It only means that such otheritems, before their disclosure in an SEC filing, are largely anticipated by the market becauseof previous information (quite possibly including a press release containing the exact sameinformation) that has become available about the corporation. An issuer that knows it willhave to make a formal filing containing certain information at specified dates will most likelydisclose more about the same subject outside the formal filings as well, and when it doesso, it will be more careful about the accuracy of what it says. To the extent that the filednumbers simply verify the investment community’s understanding of the issuer based onthis previously released information, there would be no share price reaction to these otheritems, but they well may not have been released previously but for mandatory disclosure.

is sufficient to give a sense of their magnitude. Kaplan and Reaugh conducteda survey of the 1930 annual reports of a representative sample of the nation’s500 or 600 largest publicly traded industrial corporations.78 While most firms(though not all) provided some kind of income statement, almost a third did notreveal how much depreciation, if any, was deducted to arrive at their earningsfigures.79 Without information about depreciation, earnings numbers are virtuallymeaningless.

Opponents of mandatory disclosure offer, as one of the central exhibitsin their case, Benston’s finding of no significant difference in the reduction inriskiness between firms that previously disclosed sales and firms that did not.Romano, for example, characterizes this study as “important and stillunderappreciated.”80 But Benston’s finding is surely a weak reed on which tolean. It would only take on real substance if his assumption were correct that therest of the Exchange Act’s disclosure requirements were of no importance.Benston adopts this assumption on the basis of a superficial and, as it turns out,erroneous examination of the historical state of affairs in the 1920s and 1930s.81

When his own data suggests that the assumption is incorrect, he ignores the datarather dropping the assumption or providing an alternative explanation for thedata.


82. Romano, supra note 1, at 2372. 83. Romano suggests that an increase in share price accuracy does not producebenefits. She says:

Even this alternative explanation [an increase in price accuracy] does not,however, demonstrate that the Act benefited investors. A core tenet ofmodern finance theory is that investors are compensated for bearing marketrisk, and it was firm-specific risk and not market risk that was measured tohave decreased with the 1933 Act. ... A reduction in own-return variance (thatis, more accurate stock prices) is of no value to diversified investors.Consequently, commentators who point to the return variance reduction asevidence affirming the efficacy of the 1933 Act are mistaken; investors benefitonly form reductions in risk that is priced.

Romano, supra note 1, at 2377. Romano’s statement takes an unduly narrow view of thefunctions in our economy of its securities markets. It misses the positive effects in the realeconomy, discussed in Part I, that result from more available information and more accurateshare prices. It also ignores the benefit of more accurate prices for less than fullydiversified investors, although the ability of such persons to achieve the same welfaregains by diversifying more would make this, standing by itself, a less compelling argumentfor regulation. See Fox, Shelf Registration, supra note 47, at 1009-25.

c. Conclusion. Taken as a whole, the Stigler, Simon and Benstonstudies suggest that imposition of the current system of mandatory disclosure didincrease price accuracy and the amount of meaningful information in the market.Certainly, they are at odds with Professor’s Romano’s reference to “the neartotal absence of measurable benefits from the federal regulatory apparatus.82

The evidence in these studies that mandatory disclosure does in fact enhanceprice accuracy is very important since the preceding discussion in Part I showsthat greater information availability and increased share price accuracy produceseveral social benefits: reduced risk for less than fully diversified investors,improved choice of capital utilizing projects, and reduced agency costs ofmanagement.83

Two caveats are in order, however. First, statistical comparisonsbetween time periods are tricky because other factors change besides the onemotivating the inquiry–in our case imposition of mandatory disclosure–and it maybe the changes in these other factors that cause the observed change in outcome.This should not be too great a problem here, however, since in all three studies,the measure of riskiness already factors out share price volatility caused bymarket-wide factors. Second, assuming that the imposition of mandatorydisclosure did cause the increase in price accuracy, these studies do not showwhether or not the resulting social benefits exceed the social costs of increaseddisclosure.


84. Stigler, supra note 61, at 122-124. 85. Romano, supra note 1, at 2376. 86. Irwin Friend & Fred Herman, The SEC Through a Glass Darkly, 37 J. BUS. 382 (1964). 87. George Stigler, Comment, 37 J. BUS. 414, 418-19 (1964). In the fourth year, the post-Act group did do better by a statistically significant amount. Romano argues, however,that given Friend and Herman’s lack of a theory as to why the regime should only improvethe post-Act group’s performance four years out, their recalculations do not blunt Stigler’sconclusion that Securities Act disclosure was not worthwhile. Romano, supra note 1, at2376 n.47. 88. Gregg Jarrell conducted a study making the same comparison as Stigler but withoutStigler’s methodological shortcomings. He came to the same conclusion as Stigler. GreggJarrell, The Economic Effects of Federal Regulation of the Market for New SecuritiesIssues, 24 J. LAW & ECON. 613 (1981). Jarrell’s study, however, has itself been criticized forits exclusive focus on railroads (railroads were left unregulated in the post 1933 Act period

2. Studies of the Effect of New Issue Mandatory Disclosure on Rate ofReturn.

a. Results of the Existing Studies. The most frequently cited resultsconcerning the Securities Act new issue disclosure requirement’s impact on rateof return comes from the same study by George Stigler discussed above. Stiglertook two groups of new share issues, one from the period 1923-28 (prior to thepassage of the Securities Act) and the other from the period 1949-55 (after theAct’s passage), and compared their respective five-year-post-issue growth inprices as a ratio of the growth in prices in the market as a whole. The post-Actgroup did no better than the pre-Act group.84 Professor Romano argues thatthis finding “strongly suggests that the new federal regime had, at best, no effecton investor welfare,”85 a sentiment shared by Stigler.

At the time Stigler published his results, Professors Irwin Friend andFred Herman suggested that Stigler’s results really showed that new issuemandatory disclosure was worthwhile. They argued that Stigler had madecomputational errors that understated the performance of the post-Act group.86

After recalculating, however, Stigler still found that the after-issue price growthfor a majority of the five years (including, most importantly, the fifth) was stilleither not as good as the pre-Act group or not sufficiently better than the pre-Act group to be considered statistically significant.87

Stigler’s study has also been criticized on methodological groundsbecause of a failure to account for dividends and for differences among stocksin systematic risk. Research, however, reveals no adequately conducted studythat does take account of such factors and finds, at least with respect toseasoned issuers, that the post-Act group does do significantly better statisticallythan the pre-Act group.88


so they do not really belong in a comparison example) and manufacturing (issues ofmanufacturing firms were regarded as relatively clean in the pre-Act era, whereas issues ofmining and of investment trusts, which were excluded from Jarrell’s study, were the issuesthat gave rise to most of the stories of fraud). See Smith, Comments on Jarrell, 24 J. LAW

& ECON. 677, 682 (1981); and Joel Seligman, The Historical Need for a MandatoryCorporate Disclosure System, 9 J.CORP. LAW 1, 11-12 (1983). Simon, in another study againusing the techniques of modern financial economics to consider the Stigler comparison,also agrees with Stigler’s conclusions with respect to seasoned issuers and initial publicissues of shares to be traded on the New York Stock Exchange, but finds that the pre-Actissues of shares to be traded on regional exchanges were significantly over-priced and thepost 1933 issues were not (suggesting some kind of informational inefficiency in thesemarkets that was corrected through time or by imposition of the new Act). Simon, supranote 63, at 304-308, 313. 89. Any differences that do show up between the returns enjoyed by new issueinvestors and secondary market investors are probably due either to inadequate orinaccurate risk adjustments or to information inefficiencies within the new issue market.Also historical returns – the source of the data – are only inexact (though unbiased)proxies for expected returns. The reasons why more disclosure can lead to net social gain– discussed in Part I supra – in no way depends on eliminating any such inefficiencies.In Fox, Disclosure in a Globalizing Market, supra note 14, at 2538-40 n . 78, I discuss more

b. Problems with existing studies. These studies may sound helpfulto opponents of mandatory disclosure, but they have a fundamental problem intheir design that has been missed by the prior literature in the debate. Theypurport to test whether the social benefits of mandatory disclosure exceed itssocial costs by comparing, before and after the Act, the rate of return enjoyedby purchasers of newly issued shares relative to the rate of return enjoyed byinvestors in the stock market generally. The assumption is that if the returnsenjoyed by the purchasers of newly issued shares improved relative to marketreturns generally, the Act enhanced net social welfare, and otherwise it did not.The decision to compare relative pre-Act and post-Act rates of return, ratherthan absolute ones, was probably made in order to abstract out other factorsthat affected the market generally between the two periods. Unfortunately, this“throws the baby out with the bath water.” In a well functioning capital market,there should be no difference at any point in time between the risk adjustedexpected rate of return enjoyed by new equity purchasers and that enjoyed bypurchasers of shares in the secondary market. Both before and after adoptionof the Act, the investors who purchased new share issues were free to havebeen purchasers in the secondary market instead, and vice versa.

This connection between the primary and secondary market for equitiesmeans that the ratio of risk adjusted expected returns in the two markets shouldalways be one-to-one and should be unaffected by whether the imposition ofmandatory disclosure led to a net social gain or not.89 This means that the


extensively the literature concerning the possibility of such an inefficiency in the new issuemarket and the role, if any, that greater disclosure can play in reducing it.

heated debate between proponents and opponents of mandatory disclosure asto whether this ratio increased by a statistically significant amount between thepre-Act and post-Act period has been utterly beside the point. In particular, itis invalid for Romano and Stigler to conclude that the Act did not enhance socialwelfare from the fact that the studies did not find a statistically significantimprovement in the rate of return enjoyed by the purchasers of newly issuedshares relative to the rate enjoyed by share purchasers generally.

c. The difficulty in conducting a properly designed rate of returnstudy. These problems with the existing studies raise an important question:What would be the proper way to design a test of whether imposition of newissue mandatory disclosure increased net social welfare? Start by recalling fromPart I that a disclosure-induced increase in the accuracy with which new shareissues are priced will improve project choice in the economy. Scarce capital willtherefore be better allocated. Recall that in choosing among all the proposednew real investment projects in the economy, society ideally would want toimplement them in rank order of their risk adjusted expected returns (based onall available information including what is known by the managers making eachproject proposal), with the marginal project just exhausting society’s scarcesavings for investment. Because of information asymmetries between eachproposed project’s managers and the market, the ideal will not be reached.With more accurate prices, however, fewer projects with risk adjusted expectedreturns below this marginal project will be implemented and fewer projects withrisk adjusted expected returns above this marginal project will be skipped. Asa result, there will be an increase in the number of dollars of expected future cashflow generated by projects funded through new issues of equity by a givenamount of savings. If this increase, discounted to present value, exceeds thecost of the extra disclosure, there is a net social gain.

Assume for a moment that greater disclosure does lead to such a netsocial gain. Consider the effect of this on the market for capital–the market forexpected future dollars. The same number of dollars invested today through thepurchase of newly issued equity will, with greater disclosure, generate anincreased number of expected future dollars. This increased supply drives downthe price today of an expected future dollar, whether the source of that expectedfuture dollar is newly issued equity or any other investment vehicle. Put the otherway, if greater disclosure leads to a net social gain, it leads to an increase in theeconomy-wide expected risk adjusted rate of return on investment.


90. Stigler’s study, for example, looked at new issues during the period 1949-1955 to seeif their purchasers did better than did purchasers of new issues during the pre-Act periodof 1923-1928. For the period 1949-1955, the proceeds from new stock issues constitutedonly 10.7% of the amount spent by U.S. corporations on new plant and equipment.STATISTICAL ABSTRACT OF THE UNITED STATES , Table No. 602. 91. Put another way, if increased new issue disclosure produces a net social gain, onlya small portion of it will be enjoyed by purchasers of new equity issues. The rest will beenjoyed by three other groups. One is the sellers of the shares, since the increased cashflows produced as a result of better allocation will be discounted at an only slightly higherrate. The second is investors in all the other available investment vehicles. The third is thesuppliers of all other factors of production in the economy. See Fox, Disclosure in aGlobalizing Market, supra note 14, at 2552-2570.

This suggests that the only way to determine empirically whether or notSecurities Act new issue mandatory disclosure leads to a social gain is by seeingwhether the overall risk adjusted rate of return increased after the Act wasadopted. In theory we could do this. The problem is that new stock offeringshave represented only a small fraction of the total amount of real investmentoccurring in the corporate sector each year.90 Suppose that improved disclosuresufficiently improves the choice of real investment projects funded by new issuesof equity so that, after accounting for the costs of disclosure, there is asubstantial increase in the number of future dollars expected to be generated bythe projects so chosen. In absolute dollar terms, this would represent animportant net social gain. It will represent, however, only a small percentageincrease in the number of expected future dollars from all real investmentprojects, because most real investment projects are funded in other ways. Thusthe price of a future dollar will go down only slightly, or, to state the result moreconventionally, the expected rate of return on investment, including investmentin newly issued equity, will go up only slightly.91

In sum, the proper way to design a study of the net welfare effectsarising from imposition of the Act would be to see whether there was a changein the rate of return on the overall rate of return on investment in the economygenerally. Such a study is unlikely to be provide meaningful answers, however,because the statistical power of the tests available to us are not great enough topick up the impact of the Act. Even if the Act’s requirements, by improvingallocation of funds raised by new equity issues, led to a welfare gain thatsubstantially exceeded their costs in absolute dollar terms, they would probablyhave had only a very small effect on the economy-wide risk adjusted expectedrate of return on investment. Investors in new equity issues would enjoy nogreater improvement in their expected rate of return than would any otherinvestors. Given the simultaneous changes in all the other factors that affect this


92. See Benston, supra note 65, and discussion in II.A.1.a supra . 93. The cumulative residual is a measure of how well, over a period of time surroundingthe event in question (in this case imposition of the Exchange Act) the stock outperformed(or, if negative, underperformed) what would be expected of it given general trends in themarket. See, GILSON & BLACK, supra note 23, at 193- 204, and note 65. Thus, if thecumulative residuals were positive for the nondisclosing firms, this would be affirmativeevidence that the firms were made more valuable by being forced to disclose sales,presumably because the gains from increased managerial discipline exceeded the costs ofdisclosure.

rate of return, from Federal Reserve policy to trends in the internationaleconomy, the gain that occurs as the result of imposition of new issue mandatorydisclosure would almost certainly be lost in the noise and would not show up asstatistically significant.

3. Studies of the Effect of Periodic Mandatory Disclosure on Rate ofReturn.

The only systematic inquiry into the impact of mandatory periodicdisclosure on shareholder returns is contained in the study by Benston discussedabove.92 Benston’s study is different than the studies measuring the returns tothe imposition of mandatory new issue disclosure. Those studies, as we havejust seen, are fundamentally misdesigned because they use as their measure ofthe impact of the reform an index – the ratio of the expected rate of return of theparticular securities involved to that of the market as a whole – that would notchange whatever the size and direction of the reform’s net impact on socialwelfare. Benston’s study is not vulnerable to this criticism. It compares themarket valuation of the already outstanding shares of two sets of firms, one thatwould be expected to be more affected by the reform than the other if in fact thereform, as a general matter, enhanced social welfare by reducing the agencycosts of management. The power of the statistical tests that Benston uses,however, is weak, probably unavoidably so. Benston’s study also has its ownbasic design problems. Ultimately, therefore, this study too reveals little one wayor the other about the reform’s impact on social welfare.

a. Benston’s results. Recall that Benston looked at the returns of 466New York Stock Exchange firms, 290 of which disclosed sales data beforeimposition of the Exchange Act regime (the “disclosing firms”) and 176 that didnot (the “non-disclosing” firms). He used the market model to calculate eachgroup’s cumulative average residual for the period February 1934 to June 1935,a period during which the Exchange Act mandatory disclosure requirementswere initially imposed.93 The market model is used in this fashion to control for


94. Benston, supra note 65, at 148. 95. Id. at 149. 96. Id. at 142. See supra notes 68-81 and accompanying text for a discussion of thisextremely dubious assumption. 97. The market capitalization of all NYSE listed companies at the end of 1997 wasapproximately $10 trillion. If firms representing 38% of this capitalization increased in value1/2% as a result of imposition of the disclosure requirement, this would represent a gain of$19 billion.

factors that affect returns in the market, thereby making it easier to identify theeffect, if any, of the particular event under study – in this case, imposition ofmandatory periodic disclosure. Benston found that the cumulative residualswere +.10% for the disclosing group compared with +.72% for the non-disclosing firms, thus having signs suggesting that the requirements increasedeach group’s value. Neither figure, however, was significantly different fromzero statistically.94 This leads Benston to conclude that “the disclosureprovisions of the ‘34 Act were of no apparent value to investors.”95

b. Problems with the statistical power of Benston’s tests. Benstonis overly hasty in his conclusion. The fact that the cumulative residual for thedisclosing firms is not significantly different from zero means nothing. The Actwas imposed on all NYSE companies and so any effect that it would have onthe disclosure firms would be completely absorbed by the market-wide returnthat the market model abstracts out of the cumulative residual. Benston hasalready made up his mind about these firms anyway. He starts with theassumption that imposition of the Act on them is going to have no effect becausehe considers sales to be the only important accounting number not alreadydisclosed by all NYSE firms, and the disclosing firms were already providingsales figures.96

The fact that the cumulative residual for the non-disclosing firms is notsignificantly different from zero does not mean a great deal more. The statisticalpower of the test that Benston uses is sufficiently weak that it would not pick upany plausibly sized increase in value from imposition of the sales requirement.An example will illustrate the point. Suppose that requiring firms to disclose theirsales figures had a sufficiently great positive impact on the devices that limitmanagerial discretion that, despite the costs of the requirement, it increased thevalue of the firms not previously disclosing this information by ½%. Putting thisin the context of today’s market, that would represent a net social gain of about$19 billion dollars.97 This would obviously be a huge social gain, and wouldmake requiring the disclosure highly worthwhile. Yet, due to “backgroundnoise,” the gain is not large enough, in all likelihood, to have associated with ita price change that is statistically significant.


98. The expected value of the effect on share price of these other factors is zero, but thatonly means that they are as likely to add to, as to subtract from, the effect on share priceof the imposition of the sales disclosure requirement. 99. 5.53% represents 1.96 standard deviations. If the change in value were in fact 0 andthe adjusted price change was normally distributed, 95% of the time the adjusted pricechange would be within ± 5.53%. 100. This calculation involves the distribution of possible observed values of the pricechange if the true value of the change in prices which results from imposition of the salesdisclosure requirement is an increase of 0.5%. Since the observed change in prices will beconsidered statically significant at the 95% level and have the right sign only if it is anincrease or decrease of greater than 5.53%, the question becomes: what are the chancesthat the observed change after imposition of the requirement will be of that magnitude.Because of the number of observations, the distribution of possible observed changes willapproximate a normal distribution with a mean of .5% and a standard deviation equal to thestandard error of 3.07%. The required positive change, 5.53%, is 1.80 standard deviationsabove 0.5% and so, based on standard statistical tables for the normal distribution, there

To see this, start by noting that in an efficient market, the share price ofnon-disclosing firms would have risen commensurately with the hypothetical1/2% increase in value when the Act was imposed. Other chance factors,however, almost certainly would have simultaneously affected the prices ofshares of the non-disclosure firms so that the effect of the sales disclosurerequirement on share price cannot be ascertained with certainty. Due to theseother factors, the observed change in prices, even after adjustment by themarket model, is likely to differ substantially from that 1/2% increase.98

For the price change that accompanies imposition of the sales disclosurerequirement to be considered statistically significant, it must be sufficientlydifferent from zero that one can, with reasonable confidence, reject the “nullhypothesis” that the true effect of imposing the requirement is zero and that theobserved price change results solely from the other chance factors. The“standard error” is a statistically derived estimate of the tendency of these otherfactors to cause the observed price changes to deviate from the actual effectof the imposition of the requirement on prices. The standard error in Benston’sstudy was 3.07%. As a result, the observed adjusted price change would haveto be at least 5.53% before we could reject the null hypothesis with 95%confidence, the usual standard for deeming a result “statistically significant.”99

In other words, the observed change would have to be that large in order for usto say with 95% confidence that imposition of the sales requirement had anypositive effect. There is less than one chance in twenty-five that an increaseof 1/2% in the actual value of the non-disclosure firms – a $19 billion gain – willbe accompanied by an observed change in prices sufficiently large – at least5.53% – to meet this standard.100


is a 3.6% chance that the observed change in prices will be an increase of greater than5.53%. Thus the chance that the observed change will be considered statisticallysignificant and have the correct sign is 3.6%. 101. Benston, supra note 65 at 149. To be precise, Benston said the data is“consistent” with this hypothesis. The example shows that the data is equally consistentwith the hypothesis that the disclosure provisions were of value to investors. 102. Romano, supra note 1, at 2373. 103. Id. at 2372.

This example demonstrates two things. First, it shows the danger ofconfusing statistical significance with economic significance, as Benston obviouslydid in concluding that his study showed the disclosure provisions of theExchange Act were of no apparent value to investors.101 Second, it shows theextraordinary, and unwarranted, burden that Romano puts on those who doubtthe wisdom of her proposed changes. She cites the Benston article as her primeexhibit for the proposition, stating that “there is little tangible proof of the claimthat corporate information is ‘underproduced’ in the absence of mandatorydisclosure”102 and suggesting that this “surely undermines blind adherence to thestatus quo.”103 In reality, Benston’s findings are not capable of undermininganything. Mandatory disclosure was bound to fail the test Benston set upfor it. Whatever the value to investors of the sales disclosure requirement,it was almost certainly going to appear statistically insignificant.

Like with new issue disclosure, it is unlikely that the impact of mandatoryperiodic welfare on net social welfare is ever going to be resolved empirically.Increased periodic disclosure and its consequent increase in secondary marketprice accuracy can, as discussed in Part I, reduce the extent to which managersof public corporations place their own interests above those of theirshareholders. If the value of the resulting improvement in resource allocationexceeds the cost of the disclosure, there will be a net social gain. Again, thiswould increase the number of dollars of expected future cash flow generated byexisting projects and hence increase the overall rate of return on investment.Thus, in theory, we could also determine whether Exchange Act mandatoryperiodic disclosure requirements lead to a net social gain by seeing whether thisrate increased after the requirements were initially imposed on issuers. We havethe same problem, however, as with new issue disclosure that even a substantialimprovement in absolute dollar terms would be nearly impossible to detect.

b. Design problems in Benston’s study. There are also threefundamental design problems with Benston’s use of the non-disclosing firms’average cumulative residuals as a measure of the social welfare effect of the salesdisclosure requirements. The first arises from the fact that a full 38% of the firms


104. Benston, supra note 65 at 142. 105. See supra note 66. 106. See supra notes 66, 76-79 and accompanying text.

traded on the NYSE did not disclose sales prior to imposition of the Act.104

Any impact that the Act’s sales disclosure requirement had on the value of thesefirms is therefore going to have a significant effect on the market-wide return.The market model abstracts the market-wide return out of the cumulativeresiduals. Thus the cumulative average residual of the non-disclosing firms willsignificantly understate the effect of the sales requirement on these firms and asa consequence will aggravate the problem of the weakness of the statisticalpower of the available tests.

The second design problem comes the fact that the Act not only forcedthe non-disclosing firms to begin providing their sales figures, it also forced thedisclosing firms to significantly improve the quality of the sales figures they werealready providing. Prior to the imposition of the Act, there was great variationamong these firms in how the sales figure was reported.105 Thus, if the ExchangeAct’s sales disclosure requirement increased the value of the non-disclosingfirms, it should also have increased the value of the disclosing firms, thoughperhaps by not as much. This increase in the returns of the disclosing firmswould increase the average market-wide return. Again, the market modelabstracts out the market-wide return out from each firm’s residuals, includingthose of the non-disclosing firms. Thus average cumulative residuals of the non-disclosing firms would further understate the effect of the sales requirement onthese firms and hence even further aggravate the problem of the weakness of theavailable statistical tests.

The third design problem arises from the fact that the sales disclosuresby each non-disclosure firm, as discussed in Part I, produced positiveexternalities that were enjoyed the firm’s competitors, customers andsuppliers.106 Many of the firms enjoying these externalities were disclosing firms,which, after all, constituted 62% of the firms in the study. The externalitiesenjoyed by the disclosing firms were part of the social benefits produced byimposition of the Act, but they were not captured by the average cumulativeresiduals of the non-disclosure firms since they were not reflected in the returnsof the non-disclosing firms. Indeed, these externalities will actually make theaverage cumulative residuals of the non-disclosing firms smaller. This is becausethese externalities do increase the returns of the disclosing firms and henceincrease the average market-wide return. That in turn reduces the residuals ofall the individual firms, including those of the non-disclosing firms. Thus, for aportion of the social gain arising form the sales disclosure requirement – the


107. Romano, supra note 1, at 2383-2388. 108. Romano, supra note 1, at 2362; see also Roberta Romano, THE GENIUS OF

AMERICAN LAW (1993). 109. William Cary, Federalism and Corporate Law: Reflections Upon Delaware, 83YALE L.J. 663 (1974). 110. Ralph Winter, Shareholder Protection and the Theory of the Corporation, 6 J.LEG. STUD. 271 (1977). 111. See Lucian A. Bebchuk, Federalism and the Corporation: The Desirable Limitson State Competition in Corporate Law, 105 HARV. L. REV. 1435 (1992); Merritt B. Fox, TheRole of the Market Model in Corporate Law Analysis: A Comment on Weiss and White,76 CAL. L. REV. 1015, 1035-1044 (1988); Bernard Black & Reiner Kraakman, A Self-EnforcingModel of Corporate Law,109 HA R V. L. REV. 1911, 1974-1977 (1996); William L. Cary &Melvin A. Eisenberg, CORPORATIONS: CASES AND MATERIALS 125-132 (7th ed. 1995); Ehud

positive externalities enjoyed by the disclosing firms – Benston’s test statisticwould actually produce the wrong sign: the greater the social gain, the smallerwill be the average cumulative residual of the non-disclosing firms.

B. The State Corporate Law Competition Literature

Professor Romano regards the literature concerning the social welfareeffects of competition among the states in the area of corporate law as stronglysupportive of her proposal.107 This competition develops because, underexisting choice of law rules, U.S. corporations are free to choose their state ofincorporation regardless of their physical location and where they do business.Romano characterizes this arrangement as “a responsive legal regime that hastended to maximize share value.”108

The proposition that state competition for corporate charters enhancesU.S. economic welfare is in fact a controversial one and has been the subject ofone of corporate law’s most intense debates in the last 20 years. The openingsalvo in the debate was by William Cary in an article finding the competition tobe a harmful “race to the bottom.” Cary believed that states cater to the self-interested desires of corporate managers for minimal regulation.109 Ralph Winterresponded with an argument that competition leads to a “race to the top,” sincethe state offering the corporate law rules that maximizes share value will beoffering firms the lowest cost of capital. He suggests that managers arecompelled to seek a low cost of capital by forces in both the market forcorporate control and in the market for their products.110 There are a numberof more mixed evaluations of charter competition, as well. After considering thetheoretical arguments and reviewing the empirical evidence, these studies viewcharter competition neither as so broadly helpful as Winter and Romano do, noras so broadly harmful as Cary does.111 If these more mixed evaluations are


Kamar, A Regulatory Competition Theory of Indeterminacy in Corporate Law, 98 COLUM.L. REV. (forthcoming 1998). 112. Romano, supra note 1, at 2385. 113. See supra Part IV.A.2. 114. Part I supra . Lucian Bebchuk makes the point that it is precisely becausedisclosure involves these positive externalities that makes it more suitably regulated at thefederal level rather than regulated under a system subject to competitive pressures. Hecontrasts disclosure regulation with other areas of corporate law not involving positiveexternalities, and sees these other areas as better regulated at the state level specificallybecause of the potential that creates for jurisdictional competition. Bebchuk, supra note14, at 1485-95.

correct, our experience with state charter competition provides no empiricalsupport for the proposition that issuer choice would be superior to the currentmandatory system of disclosure regulation.

Even if Romano is correct that state competition for corporate chartersis share value maximizing, Romano overstates her case when she says that “thereis no reason to expect that state competition will operate differently for securitieslaw than corporate law.”112 Unlike a firm’s decision to include, through itschoice of where to incorporate, certain corporate governance terms, a firm’sdecision to commit, through a securities regime choice, to a higher level ofdisclosure has positive externalities.113 As we have seen, this can have a crucialeffect on the workings of regulatory competition since it induces issuers todemand regimes requiring them to disclose at levels below what is sociallyoptimal.114

C. Implications of the Available Empirical Studies for Whether ReformShould be Undertaken.

The following picture emerges from the foregoing review of the availableempirical studies. There is affirmative evidence for the proposition thatmandatory disclosure has increased the amount of meaningful information in themarket and improved price accuracy. Everything else being equal, these benefitswill lead to increased social welfare. Mandatory disclosure, however, involvescosts as well. Affirmative evidence is lacking for the proposition that the benefitsare greater than the costs. Affirmative evidence is also lacking for theproposition that the costs are greater than the benefits. Given the limited powerof the statistical tests available to test these latter two propositions, mandatorydisclosure would have to have had an extraordinarily large positive or negativenet effect on social welfare for the effect to be likely to have been detected.


The empirical evidence concerning the social welfare effects ofcompetition in the state corporate law area is mixed. In any event, the evidenceis not really relevant to the debate here because the behavior regulated bycorporate law does not involve significant third-party effects and securitiesdisclosure does. Indeed, these third party effects are a primary reason for theexistence of mandatory disclosure regulation in the first place.

This picture of the available empirical studies argues against undertakingthe radical reform of issuer choice for two reasons. First, as laid out in Part I,theory suggests that there will be a substantial market failure if issuer choice isadopted. We know that at least for some additional amount of increase in thelevel of disclosure above what issuers will provide under issuer choice, theincreased benefits will exceed the increased costs. This is because under issuerchoice, each issuer will choose a regime requiring a sufficiently low level ofdisclosure that the marginal social benefit of additional disclosure exceeds itsmarginal social cost. Thus mandatory disclosure, by requiring a higher level ofdisclosure, has the unquestionable potential to increase social welfare. Theproponents of issuer choice have presented no theory as to why government isso disabled that its attempts to correct this market failure are likely to be moredamaging than the market failure itself.

Given that empirical studies have not resolved the issue one way or theother, the preponderance of our understanding of the subject is theoretical.Theory points toward retaining mandatory disclosure rather than adopting issuerchoice. Under these circumstances, it is inappropriate for the proponents ofissuer choice to argue that a lack of affirmative empirical evidence that thebenefits of mandatory disclosure exceed its costs means that we should abandonmandatory disclosure. This is particularly true given how unlikely it is that thematter will ever be capable of empirical resolution.

The second reason why this picture of the existing empirical studiesargues against adopting issuer choice rests on the prudential maxim that personsadvocating change have the burden of proof. Mandatory disclosure has beenin effect for over sixty years and is a generally well-regarded governmentprogram. To the extent that empirical studies should play a role in decidingwhether to change, it is the proponents of issuer choice that need to showempirically that mandatory disclosure does harm, not the proponents of retainingthe current system that need to show that it leads to a net social gain.

III. THE CAPACITY OF ISSUER CHOICE TO ACCOMMODATE DIFFERENCES

AMONG U.S. ISSUERS IN THEIR SOCIALLY OPTIMAL LEVELS OF DISCLOSURE

Each U.S. issuer has a socially optimal level of disclosure, where thesocial marginal benefit of the issuer’s disclosure just equals its social marginal


115. See supra I.A. 116. See Choi & Guzman, National Laws, supra note 15, at 1865-1883. 117. A more elaborate statement of this criterion follows. Regulatory competitionwould give rise to some given set of regimes from among which U.S. issuers would choose.For each issuer, there will be some regime with requirements closer to the issuer’s particular

cost.115 One issuer’s socially optimal level may well differ from another’s.Professors Choi and Guzman argue that issuer choice is desirable because itbetter accommodates these differences.116 According to them, the competitionamong jurisdictions engendered by issuer choice would lead to a range ofdifferent regimes corresponding to these differing issuer needs. Each issuerwould then choose the regime most suitable for it.

This Part appraises their argument and finds three problems. First,issuer choice may not in fact give rise to a set of regimes requiring disclosurelevels corresponding to these differing issuer needs. All the world’s majorjurisdictions may in the end require approximately the same level of disclosure.Second, even if issuer choice does give rise to a differentiated set of regimes, thepotential that this range of regimes affords for better accommodating issuerneeds will not be realized. As demonstrated in Part I, each issuer is unlikely tochoose the regime requiring it to disclose at its socially optimal level. Rather, itwill choose one requiring significantly less. Finally, Choi and Guzman offer noaccount of why, to the extent that significant differences do exist among U.S.issuers in their respective socially optimal levels of disclosure, the federalmandatory regime cannot provide different rules for different issuers. To someextent, it does so even today.

These three problems suggest that concern about accommodatingdifferences among U.S. issuers is a poor reason for adopting issuer choice.Compared to the alternatives, issuer choice is likely to lead to a greater, notsmaller, average deviation between the level that each U.S. issuer is required todisclose and the issuer’s socially optimal disclosure level.

A. The Range of Regimes Offered U.S. Issuers under RegulatoryCompetition.

1. The Need to Establish that Regulatory Competition will Lead to anAppropriately Differentiated Set of Regimes.

An essential first step in showing that regulatory competition willenhance, rather than harm, social welfare is to establish that it will give rise to aset of choices that are better tailored to the particular needs of individual U.S.issuers than the current mandatory regime.117 The proponents of issuer choice


socially optimal level of disclosure than the requirements of any other regime. Comparedto the level required by the U.S. system, the level required by this regime may be closer to,or further from, this issuer’s socially optimal level of disclosure. If on average it is further,we can confidently say, without knowing more, that issuer choice would harm socialwelfare. If on average it is closer, a necessary but not sufficient condition has been met forshowing that issuer choice improves social welfare. The inquiry then would go on towhether, for each issuer, this is the regime that the issuer would in fact choose andwhether a better course of action would be to refine the current mandatory system to takebetter account of the differences among U.S. issuers. 118. This is what Choi and Guzman loosely refer to as a “separating equilibrium.” Choi& Guzman, National Laws, supra note 15 at 1878. 119. This is what Choi and Guzman loosely refer to as a “pooling equilibrium.” Id. at1880. 120. Most of the factors that determine an individual issuer’s socially optimal level ofdisclosure are ones that U.S. issuers share more in common with each other than they dowith issuers from other countries, a point acknowledged by Choi and Guzman. Id. at 1882-83. Thus the individual levels for most U.S. issuers will be much closer to the U.S. averagethan to the world average. See infra III.B. Suppose that issuer choice results in anundifferentiated set of regimes, as the text will suggest is quite possible. See infra III.A.3. Then, assuming that the current U.S. regime at least roughly reflects this U.S. average,U.S. issuers would be better governed by it than by the regimes that would be madeavailable under issuer choice by the major jurisdictions (including the United States), whichwould instead reflect the world average. U.S. issuers that are not trying to avoid disclosurealtogether will probably feel bound to choose only from among the jurisdictions of themajor capitalist countries. Only those countries can probably provide regimes with acredible promise of expert administrative and enforcement processes.

have not demonstrated convincingly that it will. It is easy to imagine, likeChoi and Guzman, that regulatory competition would indeed lead to a range ofchoices. In this story, each jurisdiction would try to maximize the number ofissuers utilizing its regime by tailoring its requirements to appeal to a differentniche in the market.118 It is just as easy to imagine, however, that each majorjurisdiction would instead try to maximize its number of issuers by appealing tothe broadest possible segment of the market, the way the large televisionnetworks traditionally have done. This would be accomplished by offering aregime that minimizes the average distance between its requirements and thepreferences of each of the world’s issuers. In this second story, all theseregimes would have essentially the same requirements.119 U.S. issuers wouldmove from being regulated by a standard designed for the average U.S. issuerto being regulated by one designed for the average issuer worldwide. Thiswould reduce, not enhance, U.S. welfare because the only effective choices thenavailable to U.S. issuers would likely have requirements further from theseissuers’ socially optimal level of disclosure than the current U.S. mandatoryregime.120

The proponents of issuer choice see their reform as creating a marketin which jurisdictions compete to sell a product – disclosure regulation. The


121. The seminal work in this area is Harold Hotelling, Stability in Competition, 39ECON. J. 41 (1929). The modern literature in the area is surveyed in JEAN TIROLE, THE

THEORY OF INDUSTRIAL ORGANIZATION 277-303 (1988). 122. See id. at 282, 286. 123. See, e.g., Hotelling, supra note 121, at 56-57. 124. See C. Aspremont et al, On Hotelling’s S tabi l i ty in Compet i t ion, 17ECONOMETRICA 1145 (1979). 125. Choi & Guzman, National Laws, supra note 18, at 1883. 126. Choi & Guzman, Portable Reciprocity, supra note 3, at 906. 127. Id. at 917; Choi & Guzman, National Laws , supra note 15, at 1870-82.Reconstructing Choi and Guzman’s reasoning as to why they believe that a significantlydifferentiated set of choices is likely to arise from issuer choice requires putting thePortable Reciprocity and National Laws articles together. The result is not entirely

question raised by these two alternative scenarios is whether this market wouldor would not in fact make available an efficiently differentiated range of thisproduct. Product differentiation is the subject of an extensive literature inindustrial organization economics.121 This literature suggests that there is noreliable correspondence, as a general matter, between the pattern ofdifferentiation that results from market competition and what is sociallyoptimal.122 Some models suggest there will be too little differentiation, with eachproducer seeking to offer a product close to the preferences of the averageconsumer.123 Other models suggest that there will be too much differentiation,with each producer trying to soften price competition by differentiating itself fromits competitors.124

The literature on product differentiation is thus at odds with Choi andGuzman’s general suggestion that we can count on the competitive market to getright the level of differentiation:

We argue that the global securities market should be free todetermine for itself - through a market-based competitiveprocess between regimes - the amount of diversity in regimes.The market will then balance the benefit to issuers and investorsfrom multiple regimes against the cost to different countries ofmaintaining a completely different level of regulation.125

More is needed than this simple appeal to the general efficiency of marketprocesses if the proponents of issuer choice are to provide any assurance thatthe set of regimes resulting from regulatory competition will in fact improve thepossible choices available to U.S. issuers.

2. Choi and Guzman’s Approach

Choi and Guzman claim that significant differentiation is “likely.”126

While they do not lay out an explicit model on which they base this prediction,they provide a story about how a differentiated set of regimes might well arise.127


satisfactory. Their more recent piece, Portable Reciprocity, contains the conclusion thatsignificant differentiation is “likely.” Supra note 3, at 906. The rationale they give for thatconclusion, however, is simply the idea that each jurisdiction will seek to appeal to adifferent niche of issuers:

Different issuers and investors will prefer different regimes. If there is sufficient capitalmobility, therefore, competition for issuers is likely to lead to more than one regime.Countries will find themselves unable to attract all types of issuers and investors,because these securities market participants will not all seek the same regulatory regime.In response, countries will target only a part of the overall market.

Id. at 917.Choi and Guzman do not explain in this recent piece, however, why this result is more

likely than each jurisdiction seeking to appeal to the broadest segment. They instead citeto their earlier piece (National Laws, supra note 15) for “a detailed discussion of how sucha diversity of regimes may come about.” Portable Reciprocity, supra note 3, at 917 n.58.The review of their analysis in the text here is based on that earlier, more detaileddiscussion. It should be noted that in that earlier discussion, however, Choi and Guzmantake no position as to whether jurisdictional competition is more likely to lead to adifferentiated or undifferentiated set of regimes. They just note that the undifferentiatedoutcome would be unstable in the absence of each jurisdiction having some kind of naturaladvantage over certain issuers. Since each major jurisdiction almost certainly does havesuch a natural advantage today, see infra note 129, they then suggest various factors thatmight, but might not, be sufficient to counteract such a natural advantage. 128. Choi & Guzman, National Laws, supra note 18, at 1876-81.

They imagine a world with two types of issuers, A and B, and two countries, 1and 2.128 A issuers want a high level of disclosure and B issuers a low one.Each country has both types of issuers. Country 1 is large and has relativelymore A issuers. Country 2 is small and has relatively more B issuers. Initiallyeach country is isolated. Country 1's regime, reflecting the preponderance of Aissuers, requires a level of disclosure closer to A’s ideal, and Country 2's a levelcloser to B’s ideal. International capital mobility and jurisdictional competitionare then introduced. Country 2, they suggest, may seek to expand the volumeof issuers covered by its regime by moving its regime even closer to B’s ideallevel. While this may result in the loss of some A issuers, it would be more thancounterbalanced by the gain in B issuers. In response, Country 1 may resignitself to the loss of B issuers and move even closer to A’s ideal level of requireddisclosure. The final result would be a significantly differentiated set of choicesthat are closer to each type of issuer’s private optimum than prevailed onaverage before.

Choi and Guzman acknowledge that jurisdictional competition couldresult in an undifferentiated set of choices instead. They can see this happeningif Country 1 responds to Country 2's action by itself moving more toward B’sideal level in order to retain those of its B issuers that would otherwise defect toCountry 2. Through a path they do not fully describe, each country’s regimewould then end up requiring the same disclosure level. They argue, however,that unless each country has some “natural” advantage for retaining its own


129. Id. at 1879-81. Choi and Guzman’s own approach suggests a significant likelihoodthat an undifferentiated set of choices would in fact be stable. The United States and theother major capitalist countries almost certainly have, and are likely to retain for some time,a natural advantage with issuers of their own nationality. As long as financial informationis not fully globalized, U.S. investors, for example, will continue to exhibit their verysubstantial bias toward investing in U.S. issuers. See Fox, Disclosure in a GlobalizingMarket, supra note 14, at 2512-2515, 2523-29. Given the strong bias of U.S. investors forU.S. issuers, they are likely to have a strong bias toward the U.S. disclosure regime as well.Most U.S. issuers, wanting to satisfy their most natural group of investors, are thereforelikely to choose the U.S. regime. The same story can be told about issuers and investorsin each of the small number of other large capitalist countries with established regulatoryregimes.

Under these circumstances, the United States and the few other large capitalist countrieswill, according to mechanisms identified by Choi and Guzman, each seek to attract theminority of issuers of the world that do not have a strong affinity for their home country’sregime. National Laws, supra note 15, at 1880. They will do so by offering such issuers adisclosure regime reflecting the lowest common denominator, the way the large televisionnetworks traditionally competed for viewers in the United States. This will require each ofthe countries to move their regimes toward some world average required level of disclosure.

issuers, this result would be unstable. Country 2, seeing that Country 1'sabandoned A issuers are ripe for the picking, would reverse strategy and movetowards A’s ideal level of required disclosure, again creating a significantlydifferentiated set of regimes. Choi and Guzman go on to explore the extent towhich each country might in fact have such a natural advantage and the factorsthat would chip away at these advantages thereby making any temporary lackof differentiation unstable.129

3. An Alternative Approach

Consider, however, an alternative scenario that uses Choi andGuzman’s same simple assumptions and is at least as plausible. In this scenario,jurisdictional competition leads to an undifferentiated set of regimes and theresult is stable. The starting point is the same, with the same two countries, eachinitially isolated, and the same two types of issuers. Again, international capitalmobility and jurisdictional competition are then introduced. Assume that neithercountry has any natural advantage with respect to any of the issuers - the verycondition that Choi and Guzman suggest would make unstable theundifferentiated result they conjure up. Each issuer will thus choose its regimeentirely on the basis of how close each regime comes to the issuer’s privatelyoptimal level of required disclosure. Either country can move first in responseto this change in circumstances. Whichever country does will adoptrequirements equal to A’s privately optimal level since, worldwide, there are


130. A is the preponderant kind of firm in the larger country. If there were more B firmsthan A firms worldwide, the same scenario would play out, but with each jurisdictionadopting requirements equal to B’s privately optimal level. 131. The first mover can choose between (1) requiring A’s privately optimal level ofdisclosure, and (2) requiring something less. There would be no reason to require moredisclosure than A’s privately optimal level since that would be less appealing to all issuers.The second mover can choose between (1) requiring more than the first mover chose torequire (but, for the same reasons, not more than A’s privately optimal level), (2) requiringthe same level as the first mover chose, and (3) requiring less than the level the first moverchose. If the first mover chose A’s privately optimal level, the second mover’s choices (1)and (2) are the same.

Consider how this game would play out. Assume arbitrarily that there are 70 A issuersin the world and 30 B issuers. If there is a difference in the regimes’ required disclosurelevels, the A issuers choose the regime with a required level closer to their private optimumand the B issuers choose the regime closer to their private optimum. If both regimes requirethe same level, both types of issuers divide evenly between the two regimes since theissuers are indifferent between the regimes. The payoff diagram in terms of the number ofissuers adopting each regime would be as follows:

Second mover (1) (2) (3)

(1) 50,50 50,50 70,30First Mover

(2) 30,70 50,50 70,30

The dominant strategy for each jurisdiction is its choice (1). That would make it better offthan it would be under any other choice it might make given at least one of the choices ofthe other party, and at least as well off as it would be under any other choice it might makegiven any other choices of the other party.

more A firms than B firms.130 The other country responds by adopting the samestandards. This response is based on an expectation, standard in models of thistype, that if two producers offer the identical product at the same price, theywould then divide equally the available customers. In this case, Country 1 andCountry 2 would divide equally both the A issuers (who would be fullysatisfied, and indifferent as to which country to choose) and the B issuers (whowould be less than fully satisfied, but also indifferent between the two countries).The result is a stable equilibrium. This is because for each country, comparedto requiring any other level of disclosure, requiring A’s privately optimaldisclosure level would make the country better off given at least one of thechoices that the other country might make, and at least as well off given anyother choice the other party might make. In game theory terms, requiring A’sprivately optimal level of disclosure is thus the dominant strategy for bothcountries.131


132. TIROLE, supra note 121 at 286. 133. Id. 134. Id. 135. Id. 136. In explaining such behavior, they barely give mention to revenue considerations.Choi & Guzman, Portable Reciprocity, supra note 3, at 923. 137. Id. Based on a review of the literature on product differentiation, Tirole states:

It is thus clear that the incentive to differentiate products decreases

The real world, of course, is more complicated than the simpleassumptions on which any of these scenarios is based. The supply of each kindof issuer, for example, may be influenced over time by the particular level orlevels of disclosure required by the available regimes. And each jurisdiction maybe motivated in its choice of required level by more than the number of issuersit can attract to adopt its regime. As suggested by the brief discussion aboveconcerning product differentiation generally, the amount of differentiation, if any,in any given market is determined by a number of factors that pull in differentdirections. Pulling in favor of differentiation is intense price competition, sincedifferentiation tends to soften such competition.132 Pulling in favor of non-differentiation is a desire to maximize unit sales, which means choosingcharacteristics that minimize the average distance between what is offered andwhat each individual consumer wants.133 Non-differentiation is also more likelywhere a producer’s consumers gain some benefit from the mere similarity of itsproduct to that of another producer.134 The literature suggests no intuition thatone set of factors or the other should dominate as a general matter.135 In the realworld we observe both markets with a great deal of product differentiation andmarkets with almost none.

As for the particular market we are concerned with – the market fordisclosure regimes – forces for non-differentiation would, if anything, beparticularly strong. For one thing, proximity produces a clear benefit. Considerthe later of two jurisdictions to choose the requirements constituting its disclosureregime. If the later-choosing regime chooses requirements close to, rather thanfar from, those of the earlier-choosing jurisdiction, it helps investors, who havelearned to interpret the meaning of the disclosures under the earlier-choosingjurisdiction’s regime, to interpret the disclosures required under the later-choosing jurisdiction’s regime as well. Also, price competition is likely to belimited in this market. According to Choi and Guzman’s own description, unitsales maximization, not revenue maximization, appears to be the main forcedriving the behavior of jurisdictions in their choice of a required level.136 As theysee it, the larger the number of issuers choosing a jurisdiction’s regime, thegreater the size and importance of the regulating agency, the more transactionseffected on the markets located in the jurisdiction, and the greater the agency’seconomies of scale.137


when firms do not compete in prices. Indeed, Hotelling (129)enunciated the principle of minimal differentiation in suchcircumstances.

TIROLE, supra note 121, at 287. 138. It is also possible, as noted in the text, that a differentiated set of regimes willresult but that the regimes will be too differentiated from a social welfare point of view.

4. Conclusion

In sum, Choi and Guzman overstate things when they conclude that adifferentiated set of regimes is a “likely” result of jurisdictional competition. Amore appropriate conclusion would be that we have very little idea what suchcompetition would bring. Indeed, based on what we do know, anundifferentiated set of regimes seems the more likely result.138 Thus Choi andGuzman have not achieved an essential first step in demonstrating that issuerchoice will match disclosure regimes with issuers in a way that enhances, ratherthan harms, social welfare.

B. Even if a Differentiated Set of Regimes does Develop, IssuerChoice’s Capacity to Customize is not Worth its Bias toward

Underdisclosure.

Suppose that jurisdictional competition does lead to an appropriatelydifferentiated set of regimes, as Choi and Guzman hope. Thus, for each U.S.issuer, there is some jurisdiction offering a required disclosure levelapproximately equal to the issuer’s social optimum. Choi and Guzman assumethat this is the jurisdiction that the issuer will select. It is not. As discussed inPart I, the persons choosing each issuer’s disclosure regime will be itsmanagers. They have a preference for choosing a regime with a level of requireddisclosure substantially lower than the issuer’s socially optimal disclosure level.

This raises the question of whether the cure – issuer choice with itstendency toward underdisclosure – is worse than the disease – the lack ofaccommodation to individual differences among U.S. issuers under thesupposedly “one-size fits-all” federal mandatory disclosure regime. There aregood reasons to believe that the cure is worse than the disease. In other words,even if issuer choice results in an appropriately differentiated set of regimes – adubious proposition in and of itself – it may still increase the average deviationbetween each U.S. issuer’s actual level of disclosure and its socially optimal one.This is because, as discussed below, the differences in socially optimal disclosurelevels among U.S. issuers (at least among issuers that represent the bulk ofcapital in the United States) are small relative to differences between U.S.issuers and the major issuers of other countries. Thus, on the one hand, the cost


139. The internal decision-making structure arises out of a combination of the law ofthe jurisdiction of incorporation and the issuer country’s traditional business customs andpractices. Custom and practice include both the typical terms of the firm’s articles ofincorporation and how people typically behave within a given set of publicly and privatelyimposed legal constraints. The corporate finance environment is determined by a numberof factors, including the degree of concentration of share ownership, the nature of theholders of any such concentrated blocks, the rules and practices under which theseholders use their voting power singly and in cooperation with others, the extent to whichthe legal system and suppliers of finance facilitate or hinder hostile takeovers, and therelative availability of financing in different forms (equity versus debt) and from differentsources (private versus public markets). 140. See, e.g., MARK J. ROE, STRONG MANAGERS, WEAK OWNERS: THE POLITICAL ROOTS

OF AMERICAN CORPORATE FINANCE (1994); Bernard S. Black and John C. Coffee, Jr., HailBritannia? Institutional Investor Behavior Under Limited Regulation, 92 U.MICH.L.REV.1997 (1994).

of imposing a uniform U.S. system of mandatory disclosure on U.S. issuers isnot necessarily very great in terms of its lack of customization. On the otherhand, under issuer choice, issuers have a strong tendency, as we have seen, toselect regimes requiring them to disclose at a level substantially below their socialoptimums.

The effectiveness of disclosure in helping to reduce the agency costs ofmanagement and in assuring the best choice of real investment projects in theeconomy depends on an issuer’s internal decision structure and the corporatefinance environment in which the issuer operates.139 Studies in comparativecorporate governance show significant contrasts among countries in both internaldecision structures and corporate finance environments.140 Publicly tradedissuers from a given country are likely to have much more in common with eachother than with issuers from other countries in these regards and therefore inregard to the effectiveness of disclosure as well. Controlling for costs, the moreeffective disclosure is, the higher the socially optimal level of disclosure. Unlikeeffectiveness, there is no reason to expect major variations in the costs ofdisclosure across countries. Thus issuers from any given country will tend tohave optimal levels of disclosure that are closer to each other than they are tothose of issuers from other countries.

By way of illustration, a set of rough contrasts can be made between theUnited States (and Canada and, to a lesser extent, the U.K.) on the one handand Germany and Japan on the other. These contrasts suggest significantdifferences in the value of disclosure. Voting power in U.S. issuers is lessconcentrated and institutional investors in U.S. issuers are less inclined,separately or together, to exercise their voting power to influence corporate


141. See, e.g., MARK J. ROE, STRONG MANAGERS, WEAK OWNERS: THE POLITICAL ROOTS

OF AMERICAN CORPORATE FINANCE 22, 169-70 (1994). 142. E.g., F. X. Browne, Corporate Finance: Stylized Facts and TentativeExplanations, 26 APPLIED ECONOMICS 485, 488 (1994) (“[Non-financial f]irms in securities-based financial systems (the United States, the United Kingdom and Canada.) have quitelow debt/equity ratios compared to those in the bank-based systems of Japan, Germanyand France.”) 143. E.g., Id. at 494 (stating that internal funding is significantly greater in the UnitedStates, the United Kingdom, and Canada than in Japan and continental Europe). 144. Japanese firms borrow $5.33 from banks for every dollar they raise in the capitalmarkets, German firms $4.20, and American firms $0.85. Jonathan R. Macey & Geoffrey P.Miller, Corporate Governance, 48 STAN. L. REV. 73, 85, 89 (1995). 145. Ronald Gilson and Bernard Black show that the prospect of a vibrant market forinitial public offerings in the United States for issuers that have shown a certain degree ofsuccess greatly facilitates the earlier provision of venture capital to get them off the groundin the first place and explains why there is so much more venture capital available in theUnited States. Ronald Gilson & Bernard Black, Venture Capital and the Structure ofCapital Markets: Banks versus Stock Markets 47 J. FIN. ECON. 243 (1998).

decisions.141 Debt/equity ratios are lower142 and there is more use of publiclyoffered equity as a source of finance,143 particularly by relatively new companiesfinancing major projects. Hostile tender offers are more frequent, as aresolicitations of public shareholders in proxy fights. In contrast, in Germany andJapan, institutional investors play a larger role both in monitoring managerialbehavior and in supplying finance, mostly debt.144

The picture painted here suggests that the optimal level of disclosure forU.S. issuers would be higher than for German and Japanese ones. Comparedto their foreign counterparts, U.S. institutional investors do less monitoring of theway managers of issuers make both operating and project choice decisions.U.S. institutional investors collect, analyze and act on less information (bothpublic and non-public) concerning these matters. Thus more of the work ofaligning managerial and shareholder interests with respect to these decisions fallsto the devices such as the hostile takeover threat that are assisted in theireffectiveness by greater public disclosure. Greater disclosure and itsenhancement of share price accuracy is also of more assistance to good projectchoice in the United States because of the greater reliance by U.S. “start-up”companies on the public equity markets.145

It is thus apparent that among issuers worldwide, a major portion of thedispersion in their socially optimal levels of disclosure is explained by theirnationalities. This argues for an approach that applies to each issuer a regimerequiring the level of disclosure that is socially optimal for the typical issuer of itsnationality. There is certainly variation in optimal disclosure levels among U.S.issuers. This variation, however, is insufficiently great, particularly for thoserepresenting the bulk of U.S. corporate capital, to justify trying to accommodate


146. Choi & Guzman, Portable Reciprocity, supra note 3, at 923. 147. See infra Part VI.

it by a method – issuer choice – that would introduce a substantial bias towardunderdisclosure.

C. The Possibility of Providing Different Rules for Different IssuersWithin a Mandatory Disclosure Regime.

The argument that issuer choice is needed to accommodate differencesamong U.S. issuers is further undermined by the fact that an alternative exists foraccomplishing the same thing that would not involve the risk of a globallyundifferentiated set of regimes or introduce a substantial bias towardunderdisclosure. This alternative would be for the U.S. mandatory disclosureregime to provide different rules for different issuers. Less would be requiredof those with indices suggesting that their socially optimal level of disclosure isrelatively low. To some extent, the U.S. regime does so already. Small issuers,for example, can register a public offering of new shares under Regulation A,which requires less detailed disclosure. This rule appears to reflect theeconomies of scale involved in disclosure: the social benefits from disclosure riseproportionately with the size of the issue but the social costs do not. Similarly,issuers with fewer than 500 shareholders that do not trade on a national stockexchange and have not engaged in a public offering need not provide periodicdisclosure at all even if they are very large firms. This exception appears toreflect the less substantial agency costs of management in firms that haverelatively few shareholders.

Choi and Guzman need to explain why this less radical reform is not anadequate response to the problem with the mandatory system – adapting todifferences among U.S. issuers in their socially optimal levels of disclosure – thatthey give as the reason to switch to issuer choice. The closest they come todoing so is to suggest that issuer choice will “affect the incentives of domesticlawmakers to fashion regimes designed to maximize the welfare of securitiesmarket participants.”146 This whole article, however, shows that the results ofjurisdictional competition are likely to fall well short of such welfaremaximization. One could argue, of course, that the incentives of regulators in amandatory disclosure system are so perverse that its results would fall evenfarther short, but the proponents of issuer choice do not seriously attempt suchan argument.147

Choi and Guzman also suggest that issuer choice would provide issuersthe opportunity to signal their “quality” by their choice of regime:

When a firm issues in a high quality regime, investors know thatthere is relatively little risk even without examining the


148. Choi & Guzman, Portable Reciprocity, supra note 3, at 924. 149. See I.C.1., supra .

disclosures made under that regime. Therefore, the advantagesof disclosure are supplemented by the signal provided by theregime choice.148

If this were indeed an advantage of issuer choice, it is not one that could be metby even a mandatory system that attempts to differentiate among issuers. Closeexamination, however, suggests that is not such an advantage. This particularapplication of the signaling concept simply does not make much sense.According to Choi and Guzman, a “quality” issuer is apparently one with lessfirm specific (i.e., unsystematic) risk. Yet the fundamental lesson of portfoliotheory is that diversified investors care only about the systematic risk associatedwith an issuer, something about which the choice of regime signals nothing.149

Moreover, even if “quality” covers additional attributes about which investorsin fact would care, there is no special advantage in being able to ascertain it byexamining simply the issuer’s choice of regime rather than by examining all theissuer’s underlying disclosures. In an efficient market, the investor has no needto examine either, since these attributes would already be reflected in price.

IV. THE COSTS OF TRANSITION TO ISSUER CHOICE

Determining the desirability of a proposed reform requires considerationof the transition costs associated with its implementation, not just how the reformwill work in equilibrium once it has been up and running for some time. Thespecific concern here relates to currently existing, publicly traded U.S. issuers.The firms currently belonging to this group represent the bulk of productivecapacity in this country today. In the aggregate, these particular firms are likelyto continue to be a major, if diminishing, factor in the economy for severaldecades into the future. The efficient operation of these enterprises is thus amatter of vital national concern, as is the fair distribution of the wealth theygenerate. Currently, these issuers have both their new issue and periodicdisclosure regulated by the U.S. federal regime. Professor Romano andProfessors Choi and Guzman differ in how to deal with these issuers upon U.S.adoption of issuer choice, but the approach of each raises serious problems withtheir overall plan.

A. Romano’s Approach

Professor Romano does not explicitly address the question of whatwould happen to existing issuers at the time that issuer choice is adopted. Shepresumably intends that they would initially continue to be subject to the U.S.


150. Romano, supra note 1, at 2415-18. 151. See supra I.C.3. 152. Id. Every firm that engages in an IPO becomes a publicly traded issuer with itsexisting capital managed indefinitely under the discretion of management. 153. Id. 154. Id.

federal regime. Management, however, would be allowed at any pointthereafter to change regimes with the approval of a majority of theshareholders.150

If Romano’s plan is implemented, the managers of each existing issuerwill switch their firm to a regime requiring an inefficiently low level of disclosureunless the checks against such self-serving action prove strong enough to preventit. Everything else being equal, managers prefer as low a level of periodicdisclosure as possible. Low disclosure reduces the effectiveness of devices thatlimit managerial discretion and hence provides managers with more room tosatisfy their own objective functions at the expense of shareholders.151 Thispreference will exist even when the gains to the managers are smaller than thelosses to the shareholders.

In many cases the checks will indeed be insufficiently strong to preventa switch to a regime with socially suboptimal requirements. The reasons are thesame as the reasons, discussed in Part I, concerning why the market, at the timeof an IPO, will fear that an issuer’s initial choice of a regime with an efficientlyhigh level of disclosure will not endure over the whole life of the firm.152 In brief,collective action problems make the requirement of a shareholder vote anotoriously poor check on managerial preferences of this sort.153 While theswitch to a regime with an inefficiently low level of disclosure will depress shareprices in the secondary market and reduce the attractiveness of any future useof equity financing, these results do not impact managers directly. And to theextent that they impact managers indirectly, their effects are often attenuated.154

The prospect of such switches is troubling for two reasons. First, itsuggests that adoption of issuer choice is likely to lead to significant newinefficiencies involving firms that currently control the bulk of the economy’sproductive capital. Eventually, less disciplined management, combined with theinevitable turnover in the ranks of major firms generally, will cause such firms tofade in importance, but the costs during the decades-long period of transition arepotentially huge. This concern is independent of many of the problems withissuer choice identified in Parts I and II. It does not depend, for example, onthe existence of third-party effects arising from the fact that an issuer’sdisclosures are useful to other issuers and their investors. Nor is it related to therisk that a globally undifferentiated set of regimes would develop, with eachregime requiring a level of disclosure that on average is further from the socially


155. Choi & Guzman, supra note 3, at 922.

optimal disclosure level of each U.S. issuer than is the level required by thecurrent U.S. mandatory regime. Thus, even if these other problems identified inParts I and II were to prove unfounded, issuer choice’s potentially hugetransition costs make it a questionable reform.

Second, adoption of the reform could significantly redistribute wealthfrom the investors who hold the outstanding shares of existing issuers to themanagers of these issuers. These shares were initially issued and traded duringa period in which investors assumed that the managers would be bound for thelife of the firm by the disciplining effects of the current, federally mandated levelof disclosure. Issuer choice would permit the switch to a disclosure regime inwhich managers could act more to their own benefit at the expense ofshareholders. Although every legal reform alters expectations in ways thatarbitrarily redistribute wealth, we should be cautious about ones like this thatinvolve potentially major redistributions while offering at best only speculativenet gains to society. At a minimum, realizing that this redistribution will occurilluminates the large private interest in implementing issuer choice possessed bysome of its corporate management advocates.

B. Choi and Guzman’s Approach.

Under Professors Choi and Guzman’s version of issuer choice, anissuer would select a securities law regime at the time of each new issue ofsecurities.155 Unlike Romano, they propose no exit mechanism. Thus, if theirversion of issuer choice were adopted, existing public U.S. issuers wouldpresumably be deemed to have chosen the federal disclosure regime for theiroutstanding shares. They would thus continue to be bound by its periodicdisclosure requirements for the duration of their corporate lives.

Choi and Guzman’s approach thus avoids the transition costs involvedin Romano’s approach. By the same token, however, it greatly reduces thesignificance of the overall reform since, for perhaps decades, the majority ofU.S. issuers would continue to be bound by the U.S. federal system for periodicdisclosure. Moreover, for any issuer so bound, the advantages of choosing aforeign regime for a new issue of securities are diminished as well. While aforeign jurisdiction’s new issue regime might not require disclosure of certainitems that the U.S. regime does require, the issuer is generally bound in duecourse to provide these items under the U.S. periodic requirements anyway.Thus choosing a foreign regime to govern a new issue of securities at bestprovides just a bit of a delay in having to make public the items of information


156. There are, however, greater liabilities associated with providing inadequateanswers to the new issue disclosure requirements than to the periodic ones. See supranote 47. This greater potential liability may be a reason for an issuer to choose anotherjurisdiction’s regime. It also means , as a practical matter, that under the U.S. regime anissuer is likely to disclose more in response to the same questions when it is bound by thenew issuer requirements rather than by just the periodic requirements. 157. Under Securities Act Form S-3, the issuer only has to provide information aboutthe terms of the securities and the way the offering will be underwritten. Information aboutthe business of the issuer may be incorporated by reference from previous Exchange Actperiodic disclosure filings. To qualify to use a Form S-3, an issuer must have beenproviding periodic disclosure pursuant to Exchange Act requirements for at least one yearand have equity outstanding with a market value of at least $75 million. Since the bulk ofindustrial capital in the United States is controlled by the largest 500 corporations, see FOX,FINANCE AND INDUSTRIAL PERFORMANCE, supra note 47, at 117, 414-15, it is controlled byissuers that more than easily meet these requirements. 158. See Fox, Disclosure in a Globalizing Market, supra note 14, at 2610-17.

required by the U.S. new issue regime.156 Also, for most existing U.S. issuersof any significance, the effort involved in complying with the U.S. new issuerequirements may actually be lower than the effort involved with a foreignregime. Even though foreign jurisdictions require a lower overall level ofdisclosure than does the U.S. regime and little of what is asked by them issubstantially different from anything asked under the U.S. regime, the questions,and hence their appropriate answers, are somewhat different in form. Incontrast, an issuer already bound by the U.S. periodic regime that chooses theU.S. regime to govern its new issue of securities can, in most cases, meet itsrequirements simply by incorporating by reference its answers in the periodicreports it has already filed.157

If Choi and Guzman are serious about the benefits that they believe canbe achieved from a system of issuer choice, they are going to need to modifytheir proposal to more closely resemble what Romano suggests. In doing so,however, they will have to face the objection raised here to Romano’s approachthat the transition costs will be huge.

V. FOREIGN ISSUERS

Currently, a foreign issuer wishing to offer its shares in the United Statesor to have them trade there must generally comply with the requirements of theU.S. disclosure regime.158 Under issuer choice, the foreign issuer would be ableto choose its own country’s regime instead. The proponents of issuer choiceargue that this would improve international capital mobility and reduce costsbecause foreign issuers no longer would be deterred from entering U.S. marketsor seeking U.S. investors by the U.S. regime’s high disclosure requirements.These are indeed valuable benefits, but obtaining them does not require adoption


159. See Merritt B. Fox, The Political Economy of Statutory Reach: U.S. DisclosureRules in a Globalizing Market for Securities, 98 MICH. L. REV. [ ] (forthcoming 1998)(hereinafter, Political Economy); Fox, Securities Disclosure in a Globalizing Market,supra note 14, at 2618-19. Nothing in my proposal is intended to prevent a foreign issuerfrom electing to bind itself to having the U.S. regime applied to it on an ongoing basis; Ionly suggest that the U.S. regime not be mandatorily imposed on such an issuer. 160. See Fox, Disclosure in a Globalizing Market, supra note 14, at 2533-39, 2554-61.This concept that prices are discounted to reflect the investor welfare effects of theapplicable regime is also a cornerstone of the case for issuer choice. See Romano, supranote 1, at 2367; Choi & Guzman, Portable Reciprocity, supra note 3, at 925. This concepthas its limits, however. My proposal to exempt foreign issuers from the U.S. regime doesnot include, at least when the proposal is first implemented, foreign issuers engaging inIPOs in the United States and ones whose shares trade primarily in poorly developedsecurities markets such as those that exist in many of the emerging markets countries. Fox,Political Economy , supra note 159 at [ ]. The reason for this restriction is doubts aboutthe efficiency of the markets in which these securities are sold or traded. Id. Thewillingness of the proponents of issuer choice to allow such transactions without theissuers providing U.S. level disclosure is another danger of their proposal. 161. See supra I.C.2 and I.C.3. 162. See supra Parts I and II. 163. See Fox, Political Economy, supra note 159, at III.A1; Fox, Disclosure in aGlobalizing Market, supra note 14, at 2561-69.

of issuer choice with all of its associated problems. The same benefits can beobtained by maintaining the current U.S. mandatory regime but redirecting itsreach so that it applies to all U.S. issuers and to no foreign issuers (nationalitybeing determined by an issuer’s economic center of gravity).

I have argued elsewhere that redirecting the reach of the U.S. regime inthis fashion is a desirable reform.159 Briefly, my reasons are as follows. Theefficient market hypothesis assures us that an issuer’s share price will bediscounted in the market to reflect the investor welfare effects of its applicabledisclosure regime.160 This means that the primary function of disclosure ispromotion of efficiency in the real economy, not investor protection. Asdiscussed in Part I, an appropriate level of disclosure by a country’s issuers can,through its positive effects on managerial motivation and the choice of realinvestment projects, increase the returns generated by capital-utilizingenterprises.161 A competitive market in disclosure regimes will lead a country’sissuers to disclose at below that level.162 The chief losers from suboptimaldisclosure are the country’s entrepreneurial talent and labor, not the issuer’sinvestors. This is because competitive forces push capital, with its greaterinternational mobility, toward receiving a single global expected rate of return(adjusted for risk) regardless of the disclosure practices of a given country’sissuers.163

The United States thus has a strong interest in the disclosure level of allU.S. issuers. Because of the market failures involved, we would want this level


164. See Fox, Political Economy, supra note 159, at III.B. Fox, Disclosure in aGlobalizing Market, supra note 14, at 2554-61.

determined under a system of mandatory disclosure, not one of issuer choice.By the same token, the United States has little interest in the disclosure behaviorof foreign issuers, even those whose shares are sold to or traded among U.S.residents.164 There is thus no reason to mandate that foreign issuers comply withthe U.S. system. The right response to the globalizing market for securities is notto abandon the U.S. system of mandatory disclosure, but to refocus itsapplication to where it is needed. Such a response would be just as effectiveat improving international capital mobility and reducing costs as would adoptingissuer choice, but would create none of the problem of issuer choice.

VI. CONCLUSION

Issuer disclosure serves several social functions. It improves theselection of proposed real investment projects in the economy. It improves theeffectiveness of the mechanisms that help align the interests of managers andshareholders. And it reduces risk for less than fully diversified investors.However, because disclosure involves social costs, as well, there are limits onhow much of this good thing we want. Thus each U.S. issuer has an optimallevel of disclosure. The fundamental policy question is how to get each issuerto disclose at a level as close to this social optimum as possible.

One approach, tried in the United States until the Great Depression, istotal non-regulation. Each issuer is completely free to determine how much itwill disclose. There are a number of market failures associated with thisapproach that are likely to result in most issuers significantly underdisclosing.These failures arise from a number of sources. An issuer’s disclosures are usefulto its competitors, suppliers and customers, but the issuer receives no reward inreturn. The disclosures are useful as well to investors in assessing the value ofthe securities of other issuers, and again there is no reward to the issuer. Thissecond set of benefits is important because it improves both the process bywhich the interests of the shareholders and managers of other issuers are alignedand the process by which real investment projects associated with other issuersare selected. Finally, without regulation, issuer managers have trouble vouchingfor the accuracy of their disclosures or credibly committing to provide an optimallevel of periodic disclosure over time.

A second approach, used by the United States for the last 60 years, ismandatory regulation. The function of mandatory disclosure is to correct for themarket failures identified above. This second approach relies on political


processes, combined with bureaucratic expertise, to identify and enforce thesocially optimal level of issuer disclosure.

Issuer choice represents a third approach. While it somewhat improvesthe ability of issuer managers to vouch for the accuracy of their disclosures andto credibly commit to an optimal level of periodic disclosure over time, it doesnothing to correct the other two market failures associated with total non-regulation. Issuer choice also creates the danger that each major jurisdiction,in an attempt to appeal to the broadest possible segment of a global set ofissuers, will adopt the same, lowest-common-denominator required level ofdisclosure. This would be less well suited to the needs of U.S. issuers than thecurrent mandatory regime. Finally, issuer choice may involve huge transitioncosts, as the managers of many existing, publicly traded issuers may find itworthwhile to switch to inefficiently low disclosure levels.

The discussion so far – and this Article as a whole – makes out a strongcase for mandatory disclosure as the best approach. To overcome this case, theproponents of issuer choice would need to show that the governmental failuresassociated with regulation exceed the market failures likely to be associated withissuer choice. To date, they have not done so. For the most part, theseproponents assume governmental failure and ignore market failure. The crux oftheir argument for issuer choice is that the resulting competition among regulatorswill have good incentive effects. This is not a persuasive argument for changewithout a showing that there is a need for such incentives and that they are worththeir costs in terms of the market failures that issuer choice will bring.

Where might such a showing come from? It is unlikely that it will comefrom empirical studies of the effects of the imposition of the Securities Act andthe Exchange Act in the 1930s. Certainly, no study to date comes close toshowing that the governmental failures associated with these regulations weremore damaging to our economic welfare than the market failures in the periodof total non-regulation preceding these Acts. The nature of the data stronglysuggests that future empirical studies also are unlikely to resolve whether thegovernmental failure associated with mandatory disclosure is greater than themarket failure associated with issuer choice.

If a showing is to be made that the governmental failure is the larger ofthe two, it is likely that it will have to be made on the basis of theory. This is notso promising, either. Because the market failures associated with issuer choiceassure us that issuers will disclose too little, an increase in disclosure will, oversome range, increase social welfare. The proponents of issuer choice need toprovide a theory as to why government is so disabled that its attempts to correctthis shortfall are likely to require a level of disclosure outside of this range.

The literature on public choice is a possible foundation for such a theory.Indeed, some commentators have argued on just this basis that mandatory


165. Jonathan R. Macey, Administrative Agency Obsolescence and Interest GroupFormation: A Case Study of the SEC at Sixty, 15 Cardoza L. Rev. 909, 922 (1994). SusanM. Phillips & J. Richard Zecher, The SEC and the Public Interest 22-23 (1981). But seeEasterbrook & Fischel, supra note 7, at 684-85 (explicitly considering the application ofpublic choice theory to this problem but concluding mandatory disclosure should beretained). 166. For critical reactions to public choice theory see Cass R. Sunstein, Interest Groupsin American Public Law, 38 STAN. L. REV. 29 (1985); Richard H. Pildes & Elizabeth S.Anderson, Slinging Arrows at Democracy: Social Choice Theory, Value Pluralism, andDemocratic Politics, 90 COLUM. L. REV. 2121 (1990); Dorothy A. Brown, The InvisibilityFactor: The Limits of Public Choice Theory and Public Institutions, 74 Wash. U. L.Q. 179(1996). 167. See Fox, Political Economy, supra note 159, at IV.A.2.a at IV.A.2.b.iii.

disclosure represents overregulaton at the behest of the securities industry.165

There are a number of reasons to be skeptical that a successful theoreticalargument in favor of issuer choice can be built on this basis, however. To start,there is debate about the effectiveness of public choice theory in explainingregulation generally. There is hardly a consensus that most political actionconsists of self-interested rent seeking.166 Even if one believes that public choicetheory has considerable explanatory value generally, a story that the influence ofconcentrated interests have led to too high a level of mandated disclosure mayerr in not correctly identifying all of the concentrated interests involved. Somemembers of the securities industry may well desire a high level of mandatorydisclosure in order to reduce the costs of collecting information. Others,however, may prefer a low level. For example, if only a low level is required,more firms would be willing to be public companies, thereby resulting in morefee-generating initial public offerings and secondary trades. The story also omitsconsideration of the managements of established public corporations, whoseinterests are likely to favor low levels of required disclosure.167 Also, it does notaccount for the possibility that the interests of those pushing for higher disclosuremay coincide serendipitously with correction of the market failures that wouldoccur under issuer choice.

This discussion suggests that mandatory disclosure is the best of thethree approaches for getting issuers to disclose near their optimal levels. Theproponents of issuer choice have yet to make a persuasive case that theirapproach would work better, and their prospects for doing so seem dim.

university of michigan€¦ · university of michigan retaining mandatory securities disclosure:...

Documents