Copyright (c) 1989 Iowa Law Review
Iowa University

January, 1989

74 Iowa L. Rev. 351

LENGTH: 43696 words

ARTICLE: Rethinking Confidentiality

Fred C. Zacharias *

* Associate Professor, Cornell Law School. B.A. 1974, Johns Hopkins University; J.D. 1977, Yale Law School; LL.M. 1981, Georgetown University Law Center. Many friends and colleagues at Cornell and elsewhere have provided helpful comments on earlier drafts. Special thanks go to Professors Roger Cramton, Ted Eisenberg, Rob Hillman, John Leubsdorf, David Luban, Russell Osgood, Stewart Schwab, John Siliciano, and Chuck Wolfram. The Article's study, including statistical aspects and computerized data collection, could not have been completed without the help of my talented wife, Sharon Soroko, and outstanding research assistant, Marcia Kenney.

SUMMARY:
... A recent study in Tompkins County, New York, questioned laypersons and lawyers about the following attorney-client communication: ... This disclosure privilege is reflected in the everyday practice of non-legal professionals, such as physicians, who normally keep confidentiality. ... But the fairest reading of strict confidentiality rules, including the Model Rules, is that the hypothetical lawyers may not disclose the client information. ... At some level, whether to let lawyers disclose client information in particular contexts must therefore remain a normative question: Are the beneficial effects of strict confidentiality outweighed by countervailing negative effects or by societal benefits resulting from particular disclosures? Still, it seems important to try to reconcile the theory of confidentiality with the reality. ... Even more interesting were client responses to the question, "Assuming you needed the assistance of each of the following professionals, would you be more likely to give information to attorneys than to [blank]?" Table II suggests that confidentiality and what clients believe about its scope are far less important to clients than professional codes presume. ... In the hypothetical involving the client spy -- for which no exception to confidentiality even conceivably applies absent evidence of future criminal activity -- 20.8% of responding lawyers voted in favor of disclosure. ... Consider a hypothetical jurisdiction that adopts a series of exceptions to a previously strict confidentiality rule. ...

TEXT-1:
[*352] INTRODUCTION

A recent study in Tompkins County, New York, questioned laypersons and lawyers about the following attorney-client communication: n1

A client tells a lawyer the location of a "missing child" or kidnapping victim. The client is not implicated in the person's disappearance, but does not want the lawyer to disclose the information because the client "doesn't want to get involved." The client will not accept the lawyer's assurance that the lawyer could act without naming the client.

The prevailing New York confidentiality rule requires the hypothetical lawyer to remain silent. n2 Yet 72% of the lawyers responded that a "good attorney" would disclose the whereabouts of the kidnap victim. n3 Roughly 80% of the lay subjects stated that the lawyer should be allowed to disclose. n4 This Article presents the findings of the Tompkins County study and reconsiders attorney-client confidentiality in light of those findings. The Article provides substance to the notion that strict confidentiality rules go too far.

Virtually every American jurisdiction forbids attorneys to disclose client information. The professional codes vary in their edicts, ranging from nearly absolute prohibitions on attorney disclosures to general rules containing significant exceptions. n5 The rules stem from common assumptions [*353] about our legal system: clients won't confide in lawyers without confidentiality; lawyers need it to represent clients effectively.

Whatever its benefits, however, strict confidentiality also has adverse effects. Inflexibility in the rules has produced peculiarities in the law governing client secrets. n6 More important, the tradition of strict confidentiality has helped teach lawyers and clients to rationalize amoral representation. Lawyers must close their eyes to information that might prevent harm to others or that violates the lawyers' own ethical and political beliefs. Rules encouraging this persona inevitably affect lawyers' individuality and desire to consider ethics in other aspects of their practice.

Eminent commentators thus have called for empirical research testing the benefits of strict confidentiality and the validity of its justifications. n7 The academic community has, however, uniformly ignored the call. The scholarly inaction stems, perhaps, from a "feeling" that the issues are more theoretical than real. n8 For lawyers only rarely have to choose between breaching a professional code and maintaining a morally questionable silence. Through persuasion, refusal to undertake representation, or threats of withdrawal, lawyers often can convince clients to volunteer [*354] sensitive information. The dearth of reported challenges to the application of confidentiality provisions suggests that disciplinary committees may overlook well-intended violations of the letter of the rules. n9

Nevertheless, concrete situations exist in which strict confidentiality may conflict with society's interests. Academia's refusal to question and test the operation of the rules in those cases is shortsighted. At a minimum, it makes the bar look bad. n10 In contrast, basing rules or exceptions on empirically provable contentions can forestall the public perception that ethical regulations merely protect the guild.

Moreover, it is whimsical to assume that strict rules can remain free from legal attack. Forbidding lawyers to disclose information they feel morally obligated to reveal implicates serious free speech interests. Since the Supreme Court explicitly recognized lawyers' first amendment rights in the early 1970s, n11 attorneys have mounted prospective legal challenges to many speech-restrictive ethical rules that previously seemed immune. n12 A challenge to strict confidentiality is likely to turn on the nature of the empirical evidence for and against a bar's justifications for its rule. n13 The availability of relevant data will prove important to proponents and opponents of confidentiality exceptions alike. n14

Finally, a practice of allowing attorneys and discipliners to ignore strict rules in exceptional cases has costs. Compliance with the code becomes haphazard; some lawyers will feel bound by a clear disclosure prohibition, even if the bar does not intend to enforce it across the board. n15 A disciplinary policy of overlooking breaches of confidentiality misleads those clients who rely on the rules. Alternatively, observers of the bar learn that at least some lawyers are willing to disobey ethical regulations. On the one hand, that breeds distrust of lawyers. On the other, it encourages clients to [*355] expect legal assistance for their own efforts to break the law. Nonenforcement of explicit confidentiality provisions teaches lawyers to consider noncompliance with other rules that should be followed.

Because strict confidentiality rules help shape the attitudes and conduct of lawyers and clients, code drafters should confront the core situations in which secrecy may be counterproductive. n16 Whether confidentiality's systemic justifications support the inflexible rules the Model Rules propose and which exist in many states n17 is both a theoretical and empirical question. Part I analyzes confidentiality's theoretical foundation, illustrates why there is reason to question the wisdom of strict rules, and points out areas that demand factual scrutiny. n18 Part II then takes an empirical look at the interests commonly asserted in favor of the rules. It reviews one published survey concerning attorney-client privilege and presents the new Tompkins County data, which focuses specifically on confidentiality's underlying rationales. n19 The results for the first time substantiate that strict confidentiality's justifications rely upon factually vulnerable premises.

Though not conclusive, the Tompkins County study informs the debate. It suggests that looser confidentiality standards might provide benefits outweighing any costs to the adversarial system. n20 At the very least, [*356] the study results show that the critics of strict confidentiality are right to demand further empirical evaluation of the rules. Part III thus concludes by focusing on possible exceptions to confidentiality and the form future research should take. n21

I. QUESTIONING CONFIDENTIALITY RULES IN THEORY

In routine cases, attorney-client confidentiality is uncontroversial. It protects clients and benefits society by enabling the legal system to work. n22 Even when rules require lawyers to avoid disclosures society might like them to make, lawyers typically have other resources to encourage clients to act ethically. n23

The following pages, however, focus on provisions that forbid disclosure in unusual cases -- those in which independent societal interests arguably outweigh the client's interests in secrecy and the attorney cannot induce the client to disclose. The actual dilemma scenarios highlight weaknesses in confidentiality's traditional justifications that proponents of strict rules have not fully considered.

The analysis limits itself to confidentiality rules as they apply to civil representation. n24 The relationship between criminal lawyers and their [*357] clients is unique. n25 To the extent secrecy helps maintain criminal defendants' trust and contributes to quality representation, the Constitution seems to give confidentiality its blessing. n26 This Article leaves for another day the process of balancing social interests in a limited disclosure rule against criminal clients' countervailing constitutional rights. n27

[*358] A. The Justifications for Confidentiality

The primary argument in favor of attorney-client confidentiality in civil cases rests on a three-step syllogism. First, for the adversary system to operate, citizens must use lawyers to resolve disputes and the lawyers must be able to represent clients effectively. Second, attorneys can be effective only if they have all the relevant facts at their disposal. Third, clients will not employ lawyers, or at least will not provide them with adequate information, unless all aspects of the attorney-client relationship remain secret. n28 Thus, the systemic argument goes, attorney-client confidentiality is the foundation of orderly and effective adversarial justice. n29

The bar, however, has relied on other justifications for confidentiality. By encouraging clients to communicate information they would otherwise withhold from their lawyers, n30 confidentiality enhances the quality oflegal representation n31 and thus helps produce accurate legal verdicts. n32 Proponents also claim that confidentiality improves the [*359] attorney-client relationship. n33 It can foster aspects of lawyer and client "dignity." n34 And, in theory, confidentiality helps lawyers discover improprieties that the client plans, advise against them, and ultimately stop the misconduct. n35

Undoubtedly, each of these justifications has played some role in convincing code drafters to favor rules preserving secrecy. But in many ways, strict confidentiality also serves personal interests of segments of the bar. For example, the rules relieve some lawyers from the psychological costs of having to make difficult ethical decisions. Consider this hypothetical case:

An attorney obtains information from a client that would prove that another person, falsely accused of a crime, is innocent. The attorney could reveal the information without implicating the client in the crime. The client refuses to disclose the information voluntarily.

Strict confidentiality absolves the lawyer from deciding between betraying her client's (perhaps capricious) wishes and letting an innocent victim suffer.

Confidentiality rules also may benefit an attorney financially. For example:

A tax lawyer learns, from working with the books of a client, that the client has received large sums from the Russian government. The client will only say that the money is income for services rendered. The lawyer knows that the client has access to classified government documents.

A lawyer forbidden to disclose need not fear repercussions if her affiliation [*360] with the client's actions later becomes public. When questioned about the propriety of assisting the client, the attorney can hide behind the nondisclosure rules. n36 A smile or "no comment," suggesting that the questioner would act like the lawyer "if he only knew," enables the attorney to avoid the cost of bad publicity and community disapproval of her conduct. n37 In contrast, when silence subjects lawyers to accusations of wrongdoing, most codes authorize lawyers to speak. n38

Strict confidentiality provisions promote and reinforce American society's perception of lawyers as hired guns. n39 Public acceptance of the hired gun model enables the lawyers in the hypothetical cases to take on (and accept payment for) distasteful cases. n40 Indeed, the presence of confidentiality may explain why clients are willing to pay high fees to lawyers when non-lawyers might be able to provide similar services more cheaply. n41

[*361] The extent to which the profession's personal or economic interests have influenced the scope of confidentiality rules can never be known. n42 Yet their mere existence leads one to wonder whether the attorney-drafters of the strict codes -- perhaps even unintentionally -- have overemphasized the systemic justifications for confidentiality or undervalued the social benefits of less restrictive rules. n43 A code that explicitly acknowledges lawyers' right to follow their own moral instincts despite a financial risk might produce a more ethical bar that can serve society better. n44 The following sections thus evaluate the strength of the traditional justifications for strict confidentiality and the societal costs of avoiding exceptions. n45

B. Are the Traditional Justifications Sound in Theory?

At the heart of attorney-client confidentiality rules is the notion that lawyers are clients' agents, and often their fiduciaries. Agency law requires [*362] preservation of principals' confidences n46 and forbids agents to profit personally from information the principal has disclosed in secret. n47

Attorney-client rules, however, expand the responsibility of lawyers to maintain client secrets beyond agency law standards. Ordinary agents may not sell or attempt to benefit personally from disclosing confidences. But they are "privileged to reveal information confidentially acquired by [them] in the course of [their] agency in the protection of a superior interest of [themselves] or of a third person." n48 This disclosure privilege is reflected in the everyday practice of non-legal professionals, such as physicians, who normally keep confidentiality. n49 In contrast, the strict versions of attorney-client confidentiality limit an attorney's right to disclose to situations involving dangerous future crimes. n50 Not even the most liberal of lawyer codes includes a catch-all "superior interest" provision.

As a result, strict confidentiality rules forbid attorneys to disclose in a variety of situations in which other agents might have free rein to follow their consciences. In the absence of a client's declared intent to commit or participate in a crime, strict rules might well forbid an attorney to disclose a confidence to protect third parties from criminal harm, as in the kidnapping hypothetical noted in the introduction. Standard disclosure exceptions also do not cover situations in which the client plans to commit potentially tortious, but noncriminal, activity. n51 For example:

The general counsel to a firm that produces a metal alloy used in the manufacture of airplanes learns of a company study that suggests that in some high-altitude flight patterns the alloy might weaken and cause a plane to explode. The alloy does, however, meet the minimum safety standards set by the government. The [*363] lawyer urges the Board of Directors the to recall the alloy or at a minimum to inform users of its potential danger. The Board decides that the study is too inconclusive to warrant action, in light of the dire financial consequences of disclosure to company.

Unlike the doctor who must help police identify participants in crimes involving gunshots, lawyers usually must keep evidence of past criminal activity secret, n52 as in this scenario:

A client fortuitously receives an undeserved payment from the government (e.g., a duplicate welfare check or tax refund) and deposits it in a savings account. The client then contacts his attorney, who advises the client to return the money. The client refuses.

And rarely, if ever, do strict rules allow lawyers to disclose politically or morally disturbing (but noncriminal) information about their clients, such as the fact that a client organization is secretly a Nazi front.

The Appendix lists and annotates a series of hypothetical situations in which some lawyers might feel a moral compunction to disclose confidential information. In most of these admittedly exceptional scenarios, the hypothetical lawyers could attempt to convince their clients to disclose voluntarily n53 or even withdraw from the representation. But the fairest reading of strict confidentiality rules, including the Model Rules, is that the hypothetical lawyers may not disclose the client information. n54 The first step in assessing whether strict rules err in rejecting exceptions that would allow disclosure is to analyze the strength of the rules' justifications.

1. Confidentiality's Systemic Justification

To accept the modern systemic arguments in favor of confidentiality, n55 [*364] one must reach one of two conclusions: first, that clients would use lawyers significantly less if more exceptions existed; second, that clients who employ lawyers would reveal substantially less information. Both conclusions are questionable. n56

The proponents of confidentiality posit that in this complex, litigious society, our need for "trained [legal] technicians to advise men how to order their conduct" n57 militates in favor of artificial rules that encourage the use of lawyers. Yet in theory, this very need suggests that potential clients will use lawyers even if confidentiality is circumscribed. As matters become complex, laypersons have no choice but to consult the experts. The threat of being sued or the need to sue for redress of grievances necessarily drives clients to lawyers. When litigation is not involved, the inability to understand or deal with a legal matter is usually the catalyst. n58

The notion that clients may, absent the promise of full confidentiality, withhold important information from their attorneys is intuitively more palatable. n59 A client who expects the lawyer to reveal embarrassing or damaging facts may not be willing to tell all. Again, however, for the strictness of confidentiality rules to be significant in assuaging client fears, [*365] several premises about the actual practice of law must hold true. Attorneys must regularly inform clients of the rules, or clients must learn of them from independent sources. Clients must understand the explanation of confidentiality's scope. The rules must be sufficiently clear that clients can know which of their statements will remain secret. For if these premises do not hold true, hesitant clients will withhold information despite the existence of firm confidentiality guarantees.

Supporting the premises may be difficult. Even if a lawyer makes a good faith effort to explain the rules to clients, the clients are likely to remain confused at least as to details. n60 Many aspects of confidentiality are ambiguous. A few universal exceptions to confidentiality exist, n61 some of which are subject to hot debate. n62 Often the distinction between disclosable communications and secrets rests on "vague and open-textured criteria." n63 To the extent clients learn of confidentiality from sources other than their lawyers -- such as television, literature, or friends -- the explanations they receive are likely to ignore details or distinctions among the various jurisdictions' codes.

As a practical matter, clients thus probably end up with only a general understanding that attorney-client conversations usually remain confidential but occasionally may be revealed. If that is the case, creating limited [*366] additional disclosure exceptions is unlikely to affect a client's decision to confide. n64 Absent supporting empirical evidence, it is problematic to assume that clients would avoid lawyers to any significant degree merely because they cannot speak in absolute secrecy. n65

2. Enhancing the Quality of Legal Representation and Maintaining Adversarial "Truth-seeking"

Stated broadly, the claim that lawyers can be effective only when informed of all relevant facts is simply untrue. Attorneys do without information in a broad variety of contexts. n66 To make sense, the argument in support of confidentiality must thus be redefined as follows: Lawyers whose clients hide information are likely to perform less ably. By encouraging client disclosure through secrecy guarantees, the state protects clients who otherwise would jeopardize their case by withholding information.

Professor Morgan long ago questioned the need to protect uncooperative or deceitful clients. n67 The client who receives bad advice because he fails to inform the lawyer has only himself to blame. Alternatively, if the client lies to the lawyer and later finds himself confronted by the truth, the government has little reason to aid the client. The law should probably not be written for the benefit of liars or perjurers. n68

Morgan's position, however, does not do full justice to one type of client: the genuinely confused client who needs advice and representation, but unthinkingly hesitates to confide for fear his secrets will become public. Yet by definition, this category of client feels sufficiently troubled to seek legal advice. He is unlikely to undermine that advice by withholding information, particularly when told by the lawyer that full disclosure is important.

[*367] If the client does withhold particularly embarrassing items, it is not clear that the representation will be significantly affected. In some settings, lawyers actually would prefer not to be told everything the client knows. n69 Even a lawyer who ideally would like to know all relevant facts often can provide good legal advice based on partial information. Studies suggest that criminal defendants rarely are frank with their lawyers. n70 Yet the criminal justice system relies on the presumption that these clients are nevertheless fairly and well represented.

I do not suggest that confidentiality rules have no effect on client forthrightness or the quality of representation. But in the abstract, it is difficult to determine the extent of any effect. n71 If the number of clients needing and deserving the protection of absolute rules are indeed few, the interest in "assuring effective representation" may be outweighed by society's alternative interests in allowing limited disclosures. n72

3. Client Dignity and the Trust Relationship

Absolute confidentiality can enhance lawyer client relations. It often makes the client feel as if the lawyer is a true fiduciary, with loyalty to no one other than the client. It also avoids the unseemly situation in which a lawyer induces the client to be open and then informs on the client.

But these considerations alone do not justify the strictest of rules. In an ideal world, the government would promote the relationship between clients and all agents. But that does not mean it is essential to preserve confidentiality to an extreme degree. n73 Even if we accept client "autonomy" [*368] as an important value, n74 there are limits to how comfortable we want clients to be in the belief that their lawyers will never take a stand against them. n75 Arguably, client distrust will increase if the lawyer insists that she will always act in accord with the client's wishes. n76 So long as the attorney informs the client at the outset of the relationship that she may feel compelled to disclose particular types of information, subsequent disclosures are not unseemly. n77 The client may more readily accept her as an ally within the defined boundaries, both because the lawyer has exhibited integrity and because the limitations on the alliance make the total package more believable. n78

The argument that confidentiality gives "appropriate regard" to client dignity is equally vulnerable. For one, the same argument applies to all professions. More importantly, too much secrecy can be counter-productive. n79 As [*369] the Supreme Court implicitly recognized in approving a lawyer's threat to disclose a client's proposed perjury, n80 the lawyer who contributes to the notion that the client can get away with anything demeans the client as a moral individual.

4. Preventing Client Misconduct

The most appealing secondary justification for attorney-client confidentiality is that helping lawyers obtain information enables them to advise clients against committing improper acts or filing frivolous claims. n81 Yet the same empirical questions that plague the systemic justification for strict confidentiality are present here. Confidentiality probably does allow some lawyers to prevent some misconduct before it occurs. But adding limited exceptions might not substantially affect lawyers' ability to dissuade improper acts.

Moreover, it is unclear that strict confidentiality is what provokes client candor about potential improprieties. n82 In most cases, lawyers impress upon clients the importance of full disclosure to the lawyer's ability to evaluate the case. This warning alone may procure the type of information lawyers need to prevent misconduct. n83 As a factual matter, the additional disclosures strict confidentiality fosters may only marginally improve the lawyer's ability to enforce the law.

Enabling clients to discuss planned misconduct with impunity sometimes might even promote misconduct. In consulting with clients, lawyers [*370] often serve the function of psychiatrist, social worker, or priest-confessor. They provide some clients with a psychological outlet that helps the clients persist in misconduct. Empirical research might show that lawyers play this role only rarely, that the risk of promoting misconduct deserves little weight. Yet proponents of the dissuading misconduct rationale have not relied on such evidence; they do not even consider strict confidentiality's possible costs.

C. The Negative Effects of Strict Rules

Even the literature critical of strict rules has tended to focus only on the harm of secrecy to the victims and its effect on the use of lawyers. n84 No one has fully considered the negative side effects of inflexible confidentiality on the law, lawyers' attitudes, and society's approach to the profession.

1. Impact on the Substantive Law

One of the ramifications of the model codes' adoption of strict confidentiality rules was that the profession moved to protect its own interests. When the strict general prohibitions against disclosure affected the personal and economic convenience of lawyers most directly, code drafters were prompted to insert exceptions. n85 These exceptions -- now part and parcel of most strict state codes -- do not take into account clients' actual secrecy expectations. A lawyer may disclose when confidences threaten to embroil the lawyer in a future crime. n86 The lawyer may use client communications when "necessary . . . to collect his fee." n87 When the lawyer's conduct is itself called into question -- even through no fault of the client -- the lawyer also has recourse to the client's information. n88 That is so even though the lawyer could not, under most codes, use the same information to save the life or reputation of an innocent third party. n89

Perhaps because of this apparent imbalance, judges have developed a series of contorted common-law doctrines through which litigants can force disclosure of confidences under exceptions to the attorney-client [*371] privilege. n90 Yet, for the most part, the privilege exceptions do not rely on social or moral interests in disclosure. Rather, they hinge on practical and often counterintuitive notions that clients cannot anticipate when they share their secrets.

The resulting patchwork of standards governing attorney-client secrets casts doubt on the ideals to which confidentiality rules aspire. n91 The inconsistency between practice and the publicized guarantees of confidentiality may negatively affect lawyer and lay perceptions of the legal profession's ethics. Arguably, a code that approaches the need for exceptions realistically would serve clients and society in a more logical and coherent way.

[*372] 2. Impact on Lawyers' Attitudes

Although the 1908 Canons of Ethics cautioned that the attorney "must obey his own conscience and not that of his client," n92 neither modern model code assures attorney independence. n93 Their strict confidentiality rules are prime examples. They typically admit little discretion to disclose client information, even when a lawyer believes disclosure is morally required.

Nowhere has the view that lawyers must sublimate their own interests for the good of the legal system had more of an impact than on the attitudes of lawyers themselves. As Professor Flynn explains:

By . . . emphasizing only the lawyer's relation to others and the profession, the Code allows lawyers to rationalize many forms of conduct which would otherwise transgress their duties to self and, consequently, widely held moral values. The emphasis on duty to others leads naturally and dangerously to the "hired-gun" model for deciding ethical questions. n94

The traditional legal and ethical approaches to the legal profession's responsibilities tend to distort lawyers' perceptions of their own obligations as moral and autonomous individuals. n95

The hypothetical cases described earlier and in the Appendix illustrate how strict confidentiality rules can reduce the bar's ability and willingness to hear the call of morality. In each example, one would expect the lawyer to be troubled by the duty to keep silent. Ordinarily, "[d]ecisionmaking about ethical questions is and should be a bloody business, deeply dependent upon self-conception and the ability to engage in self-reflection." n96 Yet under the ethical regime in states with strict codes, the lawyer must remain silent. Counsel are encouraged to think only in terms of what the rules require, to act indifferently "to a wide variety of ends and consequences that [*373] in other contexts would be of undeniable moral significance." n97

The codes, of course, do offer lawyers a variety of practical tools with which to persuade clients to take appropriate action. n98 But when push comes to shove, if the clients insist on the status quo and confidentiality rules apply, the lawyers may not obey their consciences by disclosing information. Strict confidentiality gives lawyers little choice but to role differentiate, n99 to avoid even thinking about personal ethics. n100 When questioned about particular decisions, they can hide behind the shield of the rules. n101

In contrast, creating disclosure exceptions would force lawyers to face their dilemmas. n102 Empowering them to exercise and express moral judgments might bring lawyers and clients closer to recognizing that zealous legal representation does not necessarily mean pulling out all the stops in [*374] every case. n103 For the willingness to explore with clients the propriety of taking particular legal steps depends directly on the lawyer's willingness to examine her own actions on a normative basis.

In theory, it is obvious that inflexible ethical rules need not undermine lawyers' moral autonomy. Whatever the mandates of the professional standards, lawyers can address ethical issues and act on them according to their own value scales. n104 Any attorney has the option to engage in civil disobedience -- to violate the governing code and accept the consequences. n105 But the whole tradition of confidentiality runs in the opposite direction. Strict rules help engrain the hired gun mentality in the professional ethos. In practice, lawyers and laymen alike come to view the profession as amoral "tools" of the client, meriting the right to earn a living, but little else. n106

[*375] 3. Impact on Society's View of Lawyers

To the extent strict rules cause lawyers to engage in "moral escapism," n107 confidentiality also helps foster the public notion that lawyers lack integrity. Let us think about the matter from the clients' perspective. They are told that virtually anything they relate to the attorney will be kept a secret. If they wish to use the legal system to harass an adversary, the lawyer will discuss the possibility with them seriously. If they wish to confess a heinous misdeed they have committed in the past, they can get it off their chests without fear of reprisal. n108 The attorney will be happy to describe options, in secret, for getting around government regulations or contractual obligations -- to the point of evaluating which of the options are illegal, which are not, and which are shady but unlikely to be punished. To the extent litigation tactics are not "frivolous" or the product of bad faith, n109 the lawyer may even be willing to strategize with the client about the hardship and cost that discovery and other legal procedures can impose on an opponent and the likelihood that the burden will prompt favorable settlement.

The image of the coconspiratorial lawyer helps explain why society considers the profession unsavory. Even if the lawyer informs the client of the few exceptions to strict confidentiality rules, n110 there seems to be nothing the lawyer is unwilling to discuss behind closed doors. n111 Clients and observers of the legal profession naturally come to look upon lawyers as "dissemblers, distorters who subordinate truth to winning, and as technicians who answer to but one command, that of their client." n112

[*376] By maintaining that lawyers have no right to reveal client information even in exceptional cases, a code prevents the bar from telling the public there are certain things it will not consider doing. n113 The more professional standards prescribe conduct inconsistent with society's general ethical perceptions, the more likely society is to view the profession with cynicism. n114

II. QUESTIONING CONFIDENTIALITY RULES ON AN EMPIRICAL BASIS

The foregoing analysis illustrates the vulnerability of strict confidentiality's theoretical underpinnings. Well-respected scholars have also noted the lack of empirical evidence establishing its practical effectiveness. n115 Yet no new studies have been conducted in response. Most commentators continue simply to assert the essential nature of strict confidentiality to our legal system. n116

Perhaps implicit in the persistent reliance on intuition instead of hard data is the notion that factual information is difficult to interpret. Inevitably, any study would show some clients rely on confidentiality in the way its proponents anticipate. One then would have to confront the fabled paradox: Is the glass half empty or half full? In other words, if the paradigm of confidentiality rules holds true in a relatively small number of cases, does that support or undermine the strict rules? Would adding limited exceptions to the rules make a significant difference to confidentiality's general impact?

The following section discusses the two available sets of data, n117 one a published study on attorney-client privilege and one a more recent private survey. It attempts to lay out the significant results for future use in analyzing confidentiality rules. In some respects, the data supports confidentiality's rationales; in others, it suggests that traditional assumptions [*377] about the importance of unlimited confidentiality are wrong. As with most empirical evidence, differing interpretations are possible.

At some level, whether to let lawyers disclose client information in particular contexts must therefore remain a normative question: Are the beneficial effects of strict confidentiality outweighed by countervailing negative effects or by societal benefits resulting from particular disclosures? Still, it seems important to try to reconcile the theory of confidentiality with the reality. For the "effect" of confidentiality and its exceptions is in part a factual issue. A normative assessment of the appropriate balance between secrecy and disclosure should take into account how well strict confidentiality serves its asserted justifications and whether creating exceptions would have a significant impact.

The two studies discussed here suggest that some of the justifications commonly offered for strict confidentiality are open to question and that further empirical evidence is needed. The studies' methodologies are imperfect and their results not definitive. Yet the studies serve an important function in highlighting areas that merit further investigation and in providing insights into the avenues future empirical research should pursue.

A. The Yale Study on Attorney-Client Privilege

In 1962, the Yale Law Journal conducted a limited study of the importance and effect of attorney client privilege rules. Its primary mission was to compare the privilege accorded the bar to that granted other professions. n118 The Journal distributed questionnaires and accumulated responses from 108 laypersons, 125 lawyers, and between 12 and 51 members of several other professions, including psychology, psychiatry, social work, marriage counseling, and accounting. n119

The survey revealed widespread misinformation concerning privileges in the various professions, and particularly the attorney-client privilege. n120 Interestingly, "[l]awyers, significantly more than laymen, believe the privilege encourages free disclosure to them." n121 Seventy-one of 108 laypersons surveyed understood that, as a general matter, attorneys would not disclose confidential matter. n122 But a significant percentage of the laypersons thought that lawyers, if questioned in court, would have an obligation to [*378] reveal confidences. n123

Forty of the 108 subjects believed there should be a legal obligation of attorneys to reveal confidences when asked to disclose in court, and an additional 19 did not take a position opposing disclosure. n124 The pool split evenly on the question of whether an outright elimination of the attorney-client privilege would deter client disclosures. n125 These figures support the notion that confidentiality rules have some impact on the way clients use attorneys. But they also cast doubt on whether the effect is as substantial as proponents of confidentiality presume.

Figures comparing lay perceptions of nondisclosure principles among the different professions are even more revealing. Attorney-client confidentiality may not be as important to clients as lawyers assume. For example, laypersons seem to be more willing to speak with divorce lawyers, even without the protection of a privilege, than with marriage counselors. n126 Moreover, while a bare majority of subjects felt that the lack of a privilege would make them less willing to give information to attorneys, n127 70% believed a similar privilege existed and was necessary to induce disclosures to psychiatrists, psychologists, marriage counselors, or social workers. n128 Sixty percent of the subjects, many of whom believed an accountant-client privilege exists, n129 felt that the absence of a legal bar to disclosure would seriously hamper accountant-client discussions. n130

Several aspects of the survey are telling. First, the figures on lay perceptions of attorneys suggest that, while a preference for nondisclosure rules exists, a substantial majority of laypersons would continue to use lawyers even if secrecy were limited. Indeed, many of the subjects believed that the privilege rules should, in fact, be confined. n131 Second, the comparative figures suggest that laypersons do not perceive any dramatic contrast in the way different professionals will protect their communications. n132 What makes these results particularly significant is that most of the nonlegal professions are not governed by the same highly protective privilege that is applicable to the legal profession. n133 That the professions [*379] continue to thrive despite the "theoretical" lay hesitation to disclose absent confidentiality suggests that either consumer ignorance or a simple need for professional services is what really controls the marketplace. n134

Of course, one can not ascribe too much significance to the Yale survey's results. The study is not tailored to confidentiality per se, nor does it document how clients rely upon the governing rules and rely upon them. At the very least, lowever, the study illustrates that the degree to which antidisclosure rules encourage client candor is unclear. All of the professions considered by the survey depend on frank bilateral discussions with clients. Yet perhaps because laypersons overestimate the safety of confiding in the nonlegal professionals, the other professions operate successfully even without the legally enforceable confidentiality guarantees lawyers can offer. n135 For similar reasons, adding exceptions to make attorney-client rules consistent with those governing other professions may not devastate the bar's ability to provide its service. The Yale figures squarely call into question the need for unlimited attorney-client confidentiality rules.

B. The Tompkins County Study on Confidentiality

1. Nature of the Study

In connection with this Article, I conducted a survey of attorneys and laypersons in Tompkins County, New York. All practicing attorneys in Tompkins County received a questionnaire surveying their practices regarding attorney-client confidentiality, their understanding of confidentiality, and their perceptions of client understanding and reliance on confidentiality. Sixty-three attorneys completed the questionnaire, under guarantees of anonymity.

Simultaneously, a pool of laypersons from Tompkins County received a separate, but parallel questionnaire. A total of 105 completed the survey. The seventy-three subjects who had consulted lawyers in the past (hereinafter referred to as "clients") were asked about their experiences. The entire pool (hereinafter referred to as "laypersons") responded to questions regarding their understanding of confidentiality.

It is important to maintain an appropriate perspective on the survey results. Neither subject pool was large or diverse enough to represent the country as a whole. Although the sixty-three responding lawyers probably [*380] make up a majority of active Tompkins County practitioners, n136 the community in which they practice is rural and university oriented. Their clientele may be more personal and their subject matter less complex than, for example, large-firm urban lawyers. n137 Similarly, the lay and client pool consisted of self-selected Tompkins County residents who had volunteered to serve for fifteen dollars per day as jurors in mock trials at Cornell Law School. Many had a connection with Cornell University. By volunteering to participate in mock trials, all had exhibited an interest in legal issues. One might therefore surmise that the surveyed laypersons were more legally sophisticated than typical individual clients, n138 but not so educated as business clients one would find in commercial urban litigation. n139

The survey asked both lawyers and laypersons about a series of hypothetical disclosure situations. Building on the Yale survey, the questionnaires also inquired about the degree to which other professions abide by confidentiality rules. The responses offer insights not only into how laypersons and clients act, but also into the degree to which attorneys perceive or misperceive the importance of confidentiality to clients.

2. Identifying the Study's Focus

Several sets of responses support the proposition that some form of confidentiality rule serves confidentiality's basic rationales. Most lawyers thought confidentiality should be retained in its current form. n140 Approximately half of the lay respondents predicted that they would withhold information from attorneys if no firm obligation of confidentiality existed. n141 A substantial number of the surveyed clients claimed to have [*381] relied upon confidentiality; nearly 30% stated that they gave information to their attorneys that "they would not have given without a guarantee of confidentiality." n142

But the responses do not explain clearly why these clients relied on confidentiality. The study suggests that many clients give information not because of confidentiality guarantees, but because they view lawyers as honorable professionals who customarily promise discretion -- like doctors or accountants. If so, enabling lawyers to act like other professionals to protect "superior [third-party] interests" n143 might not have a significant effect on client trust.

The study also revealed widespread misunderstanding among clients as to the nature of confidentiality and its scope. Half of the clients who relied on confidentiality wrongly assumed that the governing standard was absolute. n144 As discussed below, this exaggeration of confidentiality may well have been prompted by common practices of the local attorneys. To the extent lawyers manipulate clients into confiding based on a mistaken view of confidentiality, that undercuts another of confidentiality's basic rationales: that confidentiality helps clients make informed choices and thus enhances their dignity and "autonomy."

The study produced similarly equivocal data concerning the claim that confidentiality helps lawyers obtain information that can be used to prevent client misconduct. The data again supports the general rationale. More than three-quarters of the surveyed lawyers stated that confidentiality at some point in their careers helped them dissuade client wrongdoing. n145 But their responses leave two ambiguities unresolved: whether the lawyers would have gotten the information from the clients without the promise of strict confidentiality; and whether that information was of the type disclosure exceptions would cause clients to hide. These questions in turn lead to the more basic issue of how well strict rules help lawyers dissuade client misconduct. For in theory, permitting disclosure would strengthen a lawyer's hand in a moral discourse with the client, enabling her to prevent more client misconduct. The Tompkins County survey suggests that under the current regime lawyers make only limited use of confidentiality's dissuasion attributes. n146

The focus of the following discussion of the Tompkins County results therefore is not on whether confidentiality as a whole comports with its general underlying theory. At least arguably, it does. Rather, the focus is on how well confidentiality serves its justifications and whether limited exceptions would undermine the rules' effects.

The study itself reveals the practical importance of considering new exceptions. For contrary to the common belief that lawyers rarely confront moral dilemmas, more than one-third of the sixty-three lawyers surveyed [*382] claimed to have asserted attorney-client confidentiality when they "personally believed that [they] should, as a good citizen, [have been] required to make information public." n147 An even higher percentage, 44.2%, have "wanted to disclose a client confidence to a governmental agency, the press, or an adversary, but refrained because of the [CPR]." n148 "Moral reasons" played a role in prompting the desire to disclose in 38.4% of the cases, "protect[ing] some aspect of public health or safety" in 38.5%, and "avoid[ing] a personal taint" in 11.5%. n149

These figures illustrate that the professional codes control when lawyers speak and implement their own moral beliefs. The survey thus highlights the importance of testing the value of strict nondisclosure rules. To do so, it is important first to understand how lawyers implement the rules and what confidentiality means to the clients they represent.

3. Lawyer Practices and Client Perceptions

Client information and understanding. If confidentiality rules are effective in producing client forthrightness, one would expect lawyers to inform clients about the existence, nature, and extent of confidentiality, preferably at the first attorney-client consultation. One would also expect clients to rely on that explanation. Perhaps the most striking revelation of the Tompkins County survey is that lawyers overwhelmingly do not tell clients of confidentiality rules. n150

Of the lawyer pool, 22.6% confessed that they "almost never" inform clients of attorney-client confidentiality; n151 59.7% stated that they inform their clients in less than 50% of their cases. n152 When the lawyers do explain confidentiality, less than half, or 42.6%, offer their explanation at the first meeting; n153 55.7% wait until the client asks, seems to hesitate to confide, or until a specific problem of confidentiality arises. n154 The true state of affairs may be even bleaker. Clients responded that 72.9% of their first attorneys [*383] never told them anything about confidentiality. n155 Only 26.5% of the client pool were told of confidentiality at their first meeting with the lawyer. n156

In light of these figures, if strict confidentiality has an impact on client behavior, clients must be learning of the rules from other sources. The survey may support that conclusion. Of the clients who said their attorney told them nothing about confidentiality, 79.1% claimed to know of confidentiality nonetheless. n157 A plurality of these clients, 41.1%, could not identify the source of their knowledge; 32.1% learned of confidentiality from friends, books, or television. n158

One of the risks of extrinsic client education is that it leads to client misunderstanding of the prevailing standards. Even well-informed third-party sources are likely to report local rules inaccurately. Literature and television are written for national audiences; their single description of confidentiality cannot differentiate among the formulations in the varying jurisdictions. As a result, one would expect readers and viewers to derive, at best, a general understanding of what confidentiality means.

The survey confirms the breadth of the Tompkins County clients' misunderstanding of the prevailing rules. The pool was asked, "When, if ever, does the law require attorneys to keep information confidential?" Only 32.8% correctly responded that lawyers must "usually" maintain confidentiality, but that "there are exceptions for particular types of information." n159 42.4% responded that confidentiality requirements are absolute. n160 The remaining 25% believed, in varying degrees, that confidentiality rules allow more liberal disclosure. n161

Extent of clients' confidence in lawyer discretion. The extent to which confidentiality rules induce full disclosure depends in part on whether clients believe lawyers follow the rules. Over 42% of the surveyed clients understood confidentiality to be absolute. n162 Yet only 19.7% believed that "attorneys as a matter of practice [always] keep information confidential." n163 Of those clients who believed in absolute confidentiality, 14.3% stated that in practice a lawyer will disclose "depending on the lawyer's personal sense of what should be kept confidential." n164

[*384] As in the Yale study, a comparison of lawyers with other professionals proved enlightening on the subject of the significance of secrecy rules to clients. Table I illustrates that lawyers, and thus perhaps attorney-code drafters, misperceive the extent to which clients trust lawyers and rely on strict rules like those effective in New York. The surveyed lawyers thought, by margins ranging from 52.3% to 100%, that clients "believe attorneys will preserve their confidences more carefully" than doctors, psychologists or psychiatrists, accountants, social workers, and realtors. In fact, a large majority of clients responded that lawyers are no more obligated to preserve confidences than doctors, psychologists, and psychiatrists; n165 a healthy minority believed the same held true for accountants and social workers. n166 A similar pattern appeared when clients were asked whether "attorneys in fact guard client confidences more carefully than" the other professionals. n167

TABLE I -- COMPARISON OF PROFESSIONS n168
% Lawyers Who
Believe Majority % Clients Who % Clients Who
of Clients Trust Think Lawyers Think Lawyers
Lawyers to Keep Must Keep Keep
Confidential Confidential Confidential
Profession More Than More Than More Than

Priests 19.0% 22.2% 14.3%
Doctors 52.3% 36.4% 35.7%
Psychologists or
Psychiatrists 64.3% 31.3% 34.5%
Accountants 86.7% 57.6% 72.4%
Social Workers 91.3% 60.3% 62.5%
Realtors 100.0% 72.7% 72.6%

Even more interesting were client responses to the question, "Assuming you needed the assistance of each of the following professionals, would you be more likely to give information to attorneys than to [blank]?" Table II suggests that confidentiality and what clients believe about its scope are far less important to clients than professional codes presume. On the whole, the clients accepted that lawyers have a higher legal obligation to preserve confidences than accountants and social workers. Yet only half of the clients were more likely to give information to attorneys. Few were prepared to trust lawyers over priests, doctors, psychologists, or psychiatrists. This data calls into question the central role the legal profession attributes to strict rules in encouraging potential clients to use lawyers and confide in them. Evidently, guaranteeing confidentiality by rule or statute [*385] is no substitute for providing quality assistance in needed services. n169

TABLE II -- COMPARISON OF PROFESSIONS II n170
% Clients Who
Think Lawyers % Clients Who Would
Must Keep Confid. Give Info. To Lawyer
Profession More Than More Readily Than To

Priests 22.2% 35.3%
Doctors 36.4% 23.9%
Psychologists or
Psychiatrists 31.3% 34.3%
Accountants 57.6% 50.7%
Social Workers 60.3% 57.4%
Realtors 72.7% 71.4%

Confidentiality's effect in inducing client communication. Despite the fact that most lawyers surveyed did not wish to see the confidentiality rules changed or limited, n171 most believed that they would get the same information from clients even if they never informed the clients about confidentiality. n172 Yet a higher percentage, 85.9%, believed they would get enough information to represent clients competently. n173 These results may be influenced by some lawyers' view that clients know of confidentiality even without being told but, the answers at least suggest that strict rules are not essential to maintaining an adversary system.

[*386] The client survey further supports this conclusion. It shows that clients never told of confidentiality may be as ready to provide information as clients who were informed. n174

The layperson survey, however, is perhaps most informative on the question of what encourages client trust: the general notion of the lawyer as a discreet professional, or the strictness of confidentiality rules. The survey asked whether the subjects would withhold information from their attorney "if [the] attorney told you that he/she could not guarantee confidentiality but that, except in unusual cases, he/she would keep information secret." A majority of laypersons answered that they would withhold information. n175 But when the same respondents were asked whether they would still withhold information if the lawyer "promised confidentiality except for specific types of information which he/she described in advance," only 15.1% said they would withhold. n176 That is not significantly different from the 11.3% of the surveyed clients who admitted to withholding information from their attorneys under current confidentiality rules. n177 These results again suggest that the general sense of trust in attorneys as professionals -- rather than particularly strict confidentiality rules -- is what fosters client candor.

4. Client Autonomy, Dignity, and the Trust Relationship

Engendering client-centered decisionmaking. The survey results raise serious questions about the degree to which confidentiality rules create a relationship of trust between the attorney and client and accord dignity to the client by enabling him to control the representation in an intelligent manner. If such purposes indeed underlie confidentiality, one would expect lawyers to offer a fairly detailed and accurate description of confidentiality. On the one hand, lawyers would tell clients what will be kept confidential. On the other, lawyers would not exaggerate confidentiality, lest they mislead clients.

Interestingly, 72.1% of the lawyers surveyed admitted that they tell their clients "only generally that all communications are confidential." n178 Only 27.8% acknowledge to their clients that any exceptions exist. n179 Of the clients surveyed, only one recalled any mention of specific exceptions. n180 The study thus suggests that, if lawyers inform their clients about confidentiality at all, they overstate its scope. n181

[*387] What makes overstatement worse is that the same lawyers know or believe that their clients misunderstand the scope of confidentiality. 41.8% of the lawyers surveyed n182 (and 48.7% of lawyers who tell their clients only generally that all communications are confidential n183 ) thought less than half their clients understand "confidentiality and its scope." Moreover, 64.8% of all lawyers n184 and 68.3% of those who overstate confidentiality n185 thought more than three-quarters of their clients believe confidentiality is absolute. These practices have had their effect. Nearly half of the clients, 42.4%, believed the law requires absolute confidentiality. n186

One can explain these figures on a practical basis. Describing confidentiality in detail might confuse clients. Telling clients of exceptions that are unlikely to come into play would deter disclosures. Explaining exceptions is, in a sense, counterproductive to the purpose of inducing clients to trust lawyers. Depending on how one views the trade-off between misleading clients and encouraging disclosure for their own good, one might even argue that lawyers should exaggerate confidentiality's scope. n187

Nevertheless, if conscious considerations drive lawyers' willingness to deceive clients, they undermine confidentiality's autonomy rationale. Lawyers hardly enhance client "dignity" as self-determinative individuals by hiding the truth from them. Exaggerating confidentiality's scope may induce clients to trust their lawyers, but it is not a healthy basis for the trust. Rather, it represents a calculated decision to encourage an inappropriate overreliance upon the lawyers' services. If subsequently discovered, that decision jeopardizes rather than enhances effective representation. n188

[*388] To the extent the Tompkins County study shows lawyers routinely use confidentiality as a lever for prying information from clients, the study supports the systemic disclosure rationale. Yet at the same time, it undermines the notion that confidentiality enhances client dignity and engenders a relationship of trust.

Public perception and the trust relationship. The Tompkins County study provides another reason to doubt the traditional assertion that confidentiality rules cause laymen to respect and trust lawyers. For it raises a basic question about how well attorneys, and attorney-code drafters, perceive public attitudes.

The study attempted to determine whether changing the codes to allow disclosures for moral and political reasons would have a beneficial or adverse effect. The survey thus asked both lawyers and laypersons to evaluate the effects of a code amendment that would allow disclosure by a lawyer who "reasonably deems it to be necessary in the public interest."

One-third of the lawyers responded that if such a change were made, "public perception of attorneys in general" would be less favorable. n189 Only 17.5% thought that public perception of attorneys would improve. n190

The corresponding layperson response illustrates that attorneys either overestimate the importance of strict confidentiality or misperceive how poorly the public currently perceives attorney morals. The lay respondents opposed the "public interest exception" for independent reasons. n191 Nevertheless, the largest group of laypersons, 44.4%, predicted that the public perception of attorneys would improve with a public interest exception. n192 The second group, 32.2%, thought that the change would have no effect. n193 Only 23.3% believed that public perception would decline. n194 Whatever the costs of allowing disclosures for moral and political reasons, a disclosure rule therefore may facilitate countervailing improvements in the public's attitude towards lawyers. n195

[*389] 5. Preventing Client Misconduct

As noted above, the Tompkins County study suggests that many lawyers use attorney-client confidentiality to obtain information that allows them to prevent potential client wrongdoing. n196 But the study also shows that lawyers pursue this avenue relatively infrequently. n197 The results are thus equivocal on how much strict confidentiality rules promote law compliance and how important that interest is.

The survey asked the lawyers what they thought "is the best reason for keeping rules protecting attorney-client confidentiality intact." Not surprisingly, the majority, 66.1%, pointed to the need "to encourage clients to discuss their cases fully." n198 But interestingly, only two lawyers, or 3.2%, justified confidentiality primarily as a means "to give attorneys an opportunity to dissuade clients from improper action." n199 The questionnaire's wording undoubtedly contributed to the low rating the preventing misconduct rationale achieved. The subjects simply may have believed encouraging client communication is more important than or is a necessary predicate for dissuading misconduct. But even excluding confidentiality's systemic justifications from the results, we are left with a third of the responses. Less than one-tenth of these selected the preventing misconduct rationale.

Several explanations are possible. Although confidentiality may help prevent client misconduct, lawyers may believe the misconduct generally would not occur in any event. n200 Alternatively, lawyers may not believe it is an important part of the lawyer's function to dissuade misconduct. Or, based on their experiences, the respondents may have determined that the need to deter client misconduct arises so infrequently that the deterrence rationale is not significant.

Whatever the reason, the preventing misconduct rationale for confidentiality does not seem to come into play on a regular basis. Lawyers themselves consider the rationale secondary. These results cast doubt on whether the preventing misconduct justification alone can support expansive, strict rules.

6. The Hypotheticals

The Tompkins County study asked all respondents to react to the extreme fact patterns listed and annotated in the Appendix. n201 The survey itself was phrased neutrally. It did not suggest any position on whether [*390] disclosure in the hypotheticals is appropriate. n202

Since the responding lawyers practiced in New York State and the lay subjects were New York State residents, their reactions are measured against New York confidentiality provisions. n203 New York's rules are among the country's strictest. The current code of professional responsibility strictly prohibits lawyers from participating in improper conduct. n204 But unlike some other jurisdictions, New York does not permit lawyers to reveal client confidences relating to noncriminal client fraud -- past or future. n205 Its code recognizes only two exceptions to confidentiality:

(C) A lawyer may reveal:

. . . .

(3) The intention of his client to commit a crime and the information necessary to prevent the crime.

(4) Confidences or secrets necessary to establish or collect his fee or to defend himself . . . against an accusation of wrongful conduct. n206

In the first five hypotheticals, involving lawyers who might wish to reveal information that could save a kidnap victim, a falsely accused defendant, a fatally ill adversary, potential passengers of a defective airplane, and potential victims of an imbalanced client, the exceptions do not apply. n207 The lawyers in the hypothetical cases may, of course, take steps to prevent the injury, including encouraging the client to avoid the harm voluntarily and even threatening to withdraw from representing the client. n208 If, however, the hypothetical clients insist that the lawyer maintain the confidence and the lawyer cannot identify additional facts to invoke section (C)(3), n209 the New York rules do not allow the attorney to disclose.

[*391] The same holds true for the second set of three hypotheticals, involving a client's sale of a house under a previous misrepresentation, retention of an undeserved government benefit, and possible participation in subornation of perjury. Some of these hypotheticals present a closer question. Arguably, the house sale and undeserved benefit cases represent ongoing crimes that might fit within the future crimes exception. As phrased, however, the hypotheticals present only past frauds. In the perjury example, a lawyer with evidence "clearly establishing" that a third party had perpetrated a fraud upon the court could disclose. n210 But New York lawyers may not use confidences to inform on a client. n211 A good lawyer no doubt would encourage the clients to correct the misconduct in all three hypotheticals, if only to protect the clients from future repercussions. n212 But the bottom line is that a New York lawyer may refuse to participate, but should not disclose.

Hypotheticals (9) through (12), political cases involving lawyers who may wish to reveal that a client is a spy, a Nazi, or a deceitful politician, also appear to be covered by the New York confidentiality rules. Under the narrow New York exception, the lawyers have insufficient evidence that the clients intend to commit a future crime. The deceitful president and vice president arguably have waived their attorney-client privilege by making statements about the subject of the representation. n213 But a parallel narrowing of confidentiality is inconsistent with the letter and spirit of the strict ethical rules. n214

In sum, the hypotheticals all present situations in which lawyers following the New York code should probably not disclose, n215 though lawyers in other jurisdictions might. n216 Yet in the Tompkins County study, the hypotheticals produced several remarkable results. First, a significant [*392] number of the lawyers surveyed either do not understand or would not follow the confidentiality rules. n217 Second, clients (and laypersons as a whole n218 ) overwhelmingly thought lawyers should be able to disclose. A substantial number of clients already assumed lawyers may disclose in the hypotheticals. n219 Only a few stated that they would be less willing to use an attorney's services if confidentiality rules allowed lawyers to disclose in these circumstances. n220

Lawyer understanding of confidentiality rules. For each hypothetical, the survey asked layers to answer the following question: "What do you think the good attorney would do?" The study allowed only three possible responses: "The attorney would disclose the information;" "the attorney would not disclose the information;" and "I really couldn't hazard a guess." It did not specifically offer the attorneys other options, such as withdrawing from the representation, but did not exclude such an alternative. The question confronting the attorneys was thus a narrow one: regardless of whatever else they might do, would a good attorney disclose or not? n221 Table III details the responses of lawyers who answered yes or no.

TABLE III -- LAWYER RESPONSES TO HYPOTHETICALS n222
Hypothetical Would Disclose Would Not Disclose

Missing kidnap victim 72.0% 28.0%
Innocent defendant 65.3% 34.7%
Fatally ill adversary 67.3% 32.7%
Defective airplane 24.0% 76.0%
Imbalanced client 35.3% 64.7%
House sale 50.9% 49.1%
Undeserved benefit 12.5% 87.5%
Perjury 10.6% 89.4%
Spy 20.8% 79.2%
Nazi 13.7% 86.3%
Vice President 26.9% 73.1%
Nixon 3.8% 96.2%

[*393] In the hypotheticals in which the rules most clearly forbid disclosure, the five danger situations, a large number of the responding lawyers thought a good lawyer would nevertheless disclose. In the kidnapping, innocent defendant, and fatally ill adversary situations, 65-72% favored disclosure. In the defective airplane and imbalanced client settings, a lesser but still significant 24% and 35.3% thought the good attorney would disclose.

In the second category of cases, involving possible past wrongdoing by the client, the responses varied. A low of 10.6% thought the good attorney would disclose in the perjury situation. As many as 50.9% anticipated disclosure in the house sale hypothetical.

The final, unsavory client situations also produced substantial variation. Perhaps on waiver reasoning, 26.9% of responding lawyers thought disclosure appropriate in the deceitful Vice President situation. Interestingly, however, only 3.8% thought a good attorney would disclose in the hypothetical concerning President Nixon, even though the same waiver reasoning would seem to apply. The Nixon case was the only hypothetical in which a negligible number of attorneys thought confidentiality rules would allow disclosure. In the hypothetical involving the client spy -- for which no exception to confidentiality even conceivably applies absent evidence of future criminal activity -- 20.8% of responding lawyers voted in favor of disclosure.

For those who believe in professional codes of conduct, the picture that results from these responses is disturbing. n223 Some attorneys might wish to make disclosures in the hypotheticals. Yet the inflexible New York rules -- presumably to encourage client forthrightness -- forbid disclosure. That a significant and often large number of New York attorneys nonetheless would disclose in all but one of the cases suggests either that attorneys are using confidentiality to mislead clients and induce disclosures under false pretenses or that attorneys themselves do not understand the rules. n224 Either scenario calls the justifications for strict confidentiality into serious question.

[*394] Client understanding of and reliance upon confidentiality. Of course, strict confidentiality's primary justifications focus on the effect of the rules on the mind-set of clients. The Tompkins County study thus tried to identify what clients think would and should happen to confidential information in extreme cases like the hypotheticals. Table IV illustrates how clients answered the question, "Under current rules, is the attorney allowed to make the facts public or tell a government agency?"

TABLE IV -- CLIENT UNDERSTANDING OF RULES IN HYPOTHETICALS n225
Hypothetical May Disclose May not Disclose

Missing kidnap victim 56.5% 43.5%
Innocent defendant 42.6% 57.4%
Fatally ill adversary 60.0% 40.0%
Defective airplane 51.7% 48.3%
Imbalanced client 28.8% 71.2%
House sale 43.9% 56.1%
Undeserved benefit 33.9% 66.1%
Perjury 48.2% 51.8%
Spy 46.3% 53.7%
Nazi 50.9% 49.1%
Vice President 46.4% 53.6%
Nixon 30.9% 69.1%

Alone, these results merely confirm other aspects of the study. Widespread misunderstanding of confidentiality is evident among clients. In 9 of the 12 hypotheticals, 40-60% of the clients believed attorneys may disclose under current confidentiality rules. At least 28.8% believed disclosure is allowed in every hypothetical case. For these clients at least, amending the rules to allow disclosure in the hypotheticals would have no adverse effect. The clients already believe the applicable exceptions exist.

The study, however, went further in attempting to ascertain the effect allowing disclosure would have on client's willingness to use lawyers. Table V shows how clients answered two relevant questions: "Should the attorney be allowed to disclose the facts?" and "If attorneys were allowed to disclose in cases such as this, would that make you less willing to use an attorney's services?"

Attorney perceptions notwithstanding, an amendment to the confidentiality rules clearly would be popular. At least 66% of clients in each of the hypotheticals believed attorneys should be able to disclose.

TABLE V -- CLIENT REACTION TO DISCLOSURE IN HYPOTHETICALS n226
Disclosure Would
Lawyer Affect Willingness
Hypothetical Should Disclose To Use Lawyer

Missing kidnap victim 84.1% 10.0%
Innocent defendant 80.0% 19.4%
Fatally ill adversary 91.2% 10.1%
Defective airplane 85.5% 14.7%
Imbalanced client 70.3% 21.2%
House sale 77.8% 13.8%
Undeserved benefit 66.7% 25.0%
Perjury 77.4% 19.0%
Spy 80.3% 14.8%
Nazi 74.6% 16.1%
Vice President 72.6% 21.0%
Nixon 67.7% 22.6%

[*395] More significant, but perhaps less conclusive, is the way clients felt allowing disclosure would affect their willingness to use attorneys. Liberalizing disclosure exceptions apparently would not have a significant impact on most clients. In each case, one-quarter or fewer responded that disclosure would cause them to be less willing to consult an attorney. n227

The significance of these responses for attorney client confidentiality rules is debatable. If adding exceptions would limit the use of attorneys by only a small population, does that small effect justify or undermine the reasoning of confidentiality rules? Because the standard rationales for confidentiality are vague, it is unclear whether the rules are written to encourage more effective use of lawyers by the many or more frequent use of lawyers by the few who hesitate. n228 Moreover, some may consider the positive effect on client candor significant; encouraging 10-25% of laypersons to use lawyers arguably marks a successful confidentiality rule, rather than a trivial one. If disclosure exceptions keep the particular clients with sensitive information from confiding, then some of the benefits of exceptions also vanish.

[*396] The Tompkins County results thus do not conclude the issue. They simply illustrate that the consequences of limiting confidentiality in the hypothetical contexts would be neither dire nor against the will of the population as a whole.

7. Conclusions

I would be the first to caution against overreliance on the Tompkins County study. Its sampling was limited, though substantial, and its methodology somewhat unscientific. n229 Nevertheless, the data provides food for thought.

The study shows that both attorneys and clients seem to misunderstand confidentiality rules. Confidentiality in general encourages client use of attorneys and client forthrightness, but perhaps not as much as proponents assume. Moreover, client reliance on confidentiality may be attributable to lawyers who overstate the scope of confidentiality or who close their eyes to client misperception of confidentiality's limits. n230

The study establishes that professional codes inhibit the speech of some lawyers who wish to disclose for reasons fitting generally within the notion of the "public interest." That, combined with the questions the study brings to light, strongly suggests further empirical research is needed before courts and code drafters can evaluate the commonly asserted and accepted justifications for inflexible confidentiality rules. As one lawyer stated, in responding honestly to the survey:

I realize that many of my answers may well reveal a profound ignorance of what the privilege [sic] is all about. I have said I see no compelling reason for change, but if, for example, the canons or other sources of ethical constraints prohibit disclosure in [some of the hypotheticals], then we should at least very carefully consider change. n231

[*397] III. LESSONS FOR FUTURE RESEARCH

A. Methodology

One implication of the Tompkins County study is that lawyers, and thus perhaps code drafters, misperceive how clients think. The bar seems to misunderstand the way clients view and rely on confidentiality; lawyers overestimate strict confidentiality's effect on the profession's public image. n232 In attempting to assess the impact of confidentiality on clients, it is therefore important for future empirical research to seek the information from laypersons and clients themselves.

That process, however, may be complicated. n233 For the Tompkins County study revealed a high degree of misunderstanding among laypersons about confidentiality's scope. To the extent possible, one must cast the empirical inquiry in terms of client actions and thought processes at the time of the actions, rather than in terms of client predictions about the effect of hypothetical rule changes.

A full study should include an interview component. Several of the Tompkins County questions were set aside because the responses reflected confusion about key terms. Other questions proved of limited usefulness because respondents were unable to answer the multiple choice questionnaire with a single response. Follow-up interviewing is the best method for parsing out respondents' multiple reasons for their actions.

An interviewing procedure would also maximize the information obtainable from subjects who have been represented by several attorneys. The Tompkins County study's written format required it to confine questions to clients' relationships with their first attorneys. That approach, while providing sound responses, prevented data from being collected about subsequent representation. A wealth of information from willing subjects thus was lost.

B. Substance

The Tompkins County questionnaire is an appropriate starting point for formulating a questionnaire. It suggests lines of inquiry that one might profitably put to more diverse subject pools. n234 The questions raised by the study also suggest numerous additional avenues worth pursuing.

Client research. The study highlighted a potential weakness in the claim that strict confidentiality rules induce client disclosures to attorneys. In particular, the study suggested that clients do not trust lawyers markedly more than other professionals. Future research would do well to focus on the question "why?"

[*398] Assuming that a larger pool of lay subjects again conclude that they trust other professionals equally, it would be relatively simple to ask for the reasons. Learning that the subjects believe nonlegal professionals are governed by equally strict legal rules would give us important information about the miseducation of clients. But it would not necessarily undermine the justifications for strict confidentiality in the legal context. Certain results, however, could call into question the need for strict rules; for example, if clients decide whether or not to confide based on the personality of the service provider rather than rules; or, if clients in fact hesitate to confide in all professionals equally, regardless of the rules.

For similar reasons, future research should isolate clients who claim to have relied on confidentiality and identify what aspects of the rules governing their attorney-client relationship were significant. Was it simply the general tone the rules set for the relationship, or were specific provisions important? The less clients understand and focus on the rules themselves, the less difference adding exceptions would make.

The primary justifications for strict confidentiality depend heavily on the notion that clients rely upon particular rules in deciding whether to confide. It is thus important for future studies to obtain a clear picture of what is in the minds of client subjects, both before speaking with their attorneys and thereafter.

The Tompkins County study, for example, suggests widespread misunderstanding of confidentiality and heavy reliance on sources other than the lawyer to explain it. n235 Future research should be more explicit. One starting point might be to establish whether (and how) clients learned of confidentiality, what exceptions, if any, they believed existed, and what role confidentiality played in their decision to confide. As with the Tompkins County survey, it would prove useful to compare that understanding with the rules that actually prevail.

This information would enable one to probe further into the effects of adopting exceptions. As in the Tompkins County study, one might ask clients hypothetically how they would react to adoption of particular exceptions. The survey format, however, risks obtaining predictions of unmeasurable accuracy. A more concrete questionnaire -- one tied closer to clients' previous representation -- can prove more valuable.

The Tompkins County questionnaire, for example, failed to identify specific confidentiality exceptions. Clients therefore were not asked whether they had information fitting within the exceptions, whether they revealed that information to their lawyers, why they confided, and whether an exception arguably covering that information would have affected their decisions. n236 A subject pool would have to be large to include a significant number of persons with information that exceptions might cover. The [*399] specific responses would, however, shed important practical light on the theoretical arguments advanced in favor of strict confidentiality rules.

Lawyer information. Initially, a diverse study should attempt to validate the important Tompkins County findings that lawyers only occasionally tell clients of confidentiality or its exceptions. Follow-up research then could focus on why lawyers do not fully inform clients of the rules.

Suppose, for example, that empirical evidence shows lawyers want clients to be fully informed, but do not mention confidentiality because they believe clients are already aware. Such responses would lead to an interesting comparison with the data collected on actual client understanding. Discovering that lawyers misperceive client understanding would, at a minimum, argue in favor of requiring lawyers to convey the substance of confidentiality rules. Coupled with information showing that clients are unaware of confidentiality but confide nonetheless, it might suggest that further confidentiality exceptions would have no significant negative impact. On the other hand, if research reveals that lawyers hide exceptions in the hope that clients will exaggerate confidentiality, code drafters would be forced to choose between conflicting confidentiality rationales: encouraging client disclosures and promoting client autonomy.

Empirical researchers might also concentrate on the significance of confidentiality's dissuading misconduct justification. The Tompkins County study suggested that many lawyers have used confidentiality in the way envisioned by proponents of this rationale. n237 Yet the study did not provide sufficient details to establish that the lawyers needed strict rules to accomplish the dissuading function. Future studies might ask a series of practical questions of lawyers who claim to have implemented the dissuading misconduct rationale: How did they get the information used to dissuade the client? Under what circumstances? What kind of information was involved? For if the information was not of the type exceptions might cover or if the lawyer could have obtained the information even if additional exceptions had existed, the Tompkins County results would not militate against loosening confidentiality rules.

Further investigation of lawyer practices could also provide important information regarding the need for inflexible rules. The Tompkins County lawyers overwhelmingly stated that they would get sufficient information to represent clients effectively without informing clients of confidentiality. n238 The ambiguity of the question underlying these responses, however, suggests additional lines of inquiry. First, regardless of whether lawyers inform clients of confidentiality, could they obtain sufficient information even if strict rules did not exist? Second, how often, in what way, and how effectively do lawyers use means other than confidentiality to encourage clients to speak? Evidence that clients respond to a lawyer's simple statement of the need for full information to help the client again would suggest that inflexible rules are not essential.

Finally, additional research into legal practice would help define the types of exceptions code drafters might consider. The Tompkins County [*400] lawyers professed a high incidence of wanting or feeling morally obligated to disclose client information. Knowing what types of situations have arisen -- factually rather than hypothetically -- would help drafters understand the range of problems that need to be addressed. Exceptions can allow disclosure, require it, or authorize the lawyer to seek permission from a court to breach confidence. Asking lawyers who faced actual dilemma situations in the past how they would have acted under each version of an applicable exception would clarify the effects of different formulations.

C. What Exceptions Might Future Research Consider?

Aside from the traditional future crimes exception to confidentiality, different jurisdictions have allowed lawyers to disclose fraud, n239 a danger of substantial bodily harm, n240 and past client improprieties that have involved the lawyer's services. n241 Exceptions in theory also could distinguish among clients (e.g., criminal defendants, corporations) or subject matters (e.g., matrimonial, real estate, tax).

This Article does not attempt to suggest, describe, or evaluate the almost infinite variety of exceptions states might adopt. It is, however, important to note that the Tompkins County laypersons distinguished among exceptions in predicting how changes in confidentiality rules would affect their conduct. Future researchers must be specific.

To aid any efforts to delineate possible exceptions, the following pages categorize the normative reactions of the Tompkins County subjects. Numerous hypotheses can explain each set of responses. n242 The categorizations merely identify possible distinctions the subjects might have drawn. These in turn suggest varieties of disclosure exceptions future researchers should consider investigating.

Harm to third parties. The study posed a series of hypotheticals in which a lawyer potentially could use confidential information to prevent harm to third parties. Confidentiality exceptions might categorize harm situations according to the likelihood of the harm, its magnitude, or its nature. The range of opinions by the Tompkins County subjects on whether lawyers should disclose in the different hypothetical situations suggests that clients and lawyers may attach differing significance to each factor.

The Appendix includes four cases in which the hypothetical lawyer has information enabling him to prevent likely harm to specific third parties. n243 Clients overwhelmingly stated that the lawyer should be allowed [*401] to disclose, in percentages ranging from 77.8% to 86.4%. n244 A majority of the responding lawyers agreed that a good lawyer would disclose in each case, regardless of the prevailing ethical rules. The lawyers, however, seemed to draw a distinction between cases involving physical harm and the one case involving financial harm. At least 65% believed lawyers should disclose to assist a kidnap victim, an innocent defendant, and a fatally ill adversary. But only 51% believed the financial burden a third party would suffer from the client's past fraud merited disclosure. The client pool did not draw a similar financial/physical harm distinction.

Clients seemed more ready than lawyers to allow disclosure of information about client conduct that might prevent speculative harm to unknown third parties. Over 85% believed a lawyer should be able to disclose the possibility that an alloy used in airplane construction might be dangerous; 70.3% would have authorized a lawyer to reveal that a mentally unbalanced client might hurt someone. n245 In contrast, far less than a majority of lawyers (24% and 35.3%, respectively) felt that a good lawyer would disclose such speculative information.

Other aspects of the hypotheticals also suggest that clients worry about the degree of possible harm while lawyers are more concerned with being sure harm will result. The two hypotheticals that presented perhaps the greatest danger of widespread harm were hypothetical (4), the airplane case, and hypothetical (9), in which a lawyer has unconfirmed reasons to believe his client is a spy with access to classified documents. In the airplane case, a majority of clients stated that the hypothetical lawyers are and should be able to reveal the admittedly speculative danger. n246 The figures declined only slightly with respect to the suspected spy: 46% believed lawyers may reveal while 80.3% believed disclosure is appropriate. n247 In contrast, less than a quarter of the lawyers thought a good lawyer would disclose. n248

Harm to the state. Almost by definition, cases involving harm to the state usually involve some form of past client misconduct or unethical behavior. To the extent a client consults a lawyer for representation as to a past crime, the core rationales of confidentiality apply. The Tompkins County study thus did not pose any hypotheticals involving such cases.

Four hypotheticals, however, concern peripheral situations. In hypothetical (9), involving the Russian spy, and hypothetical (8), in which a friend of an organized crime client may have suborned perjury, the lawyer has a strong but unproven suspicion that his client has committed a criminal act not the direct subject of the representation. In hypothetical (7), the lawyer knows the client has fraudulently accepted a government benefit and is unwilling to purge himself of the fraud. In hypothetical (12), involving President Nixon, only unethical conduct is at issue; the lawyer knows that the president has lied publicly about his activities in office.

[*402] Unlike the third-party harm situations, most of the surveyed lawyers accepted that a good lawyer would remain silent in these cases. In the hypotheticals in which potential "harm to the state" was concerned, the lawyers did not recognize a higher law overriding the New York code's prohibition against disclosure. The client pool, on the other hand, envisioned a greater responsibility of lawyers to take steps against possible criminal conduct in the spy and perjury situations. n249 Yet the clients seemed to distinguish the need for disclosure concerning the retention of the undeserved government benefit and the President's public lie.

Two possible reasons for these distinctions come to mind. Clients may, as in the third-party harm context, believe disclosure is more appropriate the greater the potential harm. Or, they may view the latter situations like typical criminal cases, in which the harmful effects of the client conduct are largely past. Either approach easily could be incorporated into a rule authorizing disclosures of some client information.

Fraud. Three of the hypotheticals concern a form of fraud. The house sale in hypothetical (6) involves a past fraud upon a third party. The client who retains an undeserved benefit check in hypothetical (7) defrauds the government. Suborning perjury in the organized crime example of hypothetical (8) may constitute a past or future fraud upon a court.

As noted above, client fraud is not a grounds for breaching confidence in New York state. n250 But, one could, in theory, allow lawyers to prevent fraud and limit such an exception in terms of the victim, the fraud's timing, or the nature of the fraud. The Tompkins County clients, however, drew little distinction among the three hypothetical cases. Over two-thirds favored disclosure in each fraud situation -- whether past, present, or future and whether upon individuals or the state. n251

On the other hand, only the house sale example prompted a majority of lawyers to say a "good lawyer" would disclose. n252 The lawyer reactions may have stemmed from an interpretation of the house sale situation as a future crime scenario fitting within New York's disclosure exception. Alternatively, the lawyer subjects may have distinguished on the grounds that disclosure is more important when a client injures a private party. Or, they may have felt disclosure was justified when a lawyer cannot, in continuing the representation, avoid taking advantage of the fraud; arguably, in the perjury context the hypothetical lawyer could avoid use of the perjured witness's testimony.

Politically or morally distasteful clients. In several of the hypotheticals the lawyer, after agreeing to represent a client, learns facts that give the lawyer a moral or political reason to disassociate himself from and disclose secret facts about the client. Hypotheticals (11) and (12) involve public officials [*403] who have publicly stated mistruths about their previous conduct. Hypothetical (10) presents a situation in which a lawyer learns his client's organization is a Nazi front.

Most of the lawyers stated that the good lawyer would abide by strict confidentiality rules. n253 Apparently, these subjects believed that a lawyer's ability to withdraw from representation is adequate to satisfy the moral and political interests.

Many in the client pool again disagreed. Even though only 30.9% believed the prevailing rules allow disclosure in the Nixon hypothetical, 67.7% felt the lawyer should be able to disclose. In both the vice president and Nazi examples, over 72% believed the lawyer had an obligation to set the public record straight. n254

The mechanism for disclosure. In framing future research, the format of potential confidentiality exceptions will be significant. Tompkins County lay subjects sent a strong signal that vague exceptions would affect their willingness to confide far more than specific exceptions. n255 Only 15.1% felt the same way about limited, specific exceptions. n256

The effects of vagueness can result from a code's procedure for disclosure as well as its substance. For example, rules can authorize disclosure in an attorney's discretion or make disclosure mandatory. n257 A discretionary approach, by definition, would lead to fewer disclosures than the mandatory. Yet, it might have encouraging side effects in forcing lawyers to engage in moral self-examination. On the other hand, according lawyers discretion reduces clients' ability to predict the consequences of confiding. This consequence would be of particular significance to code drafters concerned with promoting client autonomy.

More radical formats are of course possible. Codes could redefine the role of lawyers, categorizing them as traditional agents and prescribing confidentiality according to agency law. Such an approach would forbid lawyers to disclose against a client's interests for personal gain, but would in practice create a vague, discretionary exception to protect third parties [*404] from harm. It might accord with what the past studies suggest are clients' general perceptions of lawyer trustworthiness; that is, that lawyers keep confidences only to the same extent as other trusted service providers. n258

Code exceptions might require lawyers planning disclosure to seek leave of court or to take other steps short of disclosure initially. Professors Shaffer and Pepper suggest that most ethical dilemmas would resolve themselves if lawyers engaged in more frequent "moral dialogues" with their clients. n259 Code drafters easily could require lawyers to consult with their clients before implementing a disclosure exception. Such protections might alleviate the hesitation of clients to confide in their attorneys. n260 By the same token, however, they impose obstacles to disclosure and thereby discourage it even in extreme situations.

The above discussion suggests that an exception's format should reflect a balance of the general confidentiality rule's primary purposes and the reasons for creating the exception. Unfortunately, code drafters traditionally have refused to define their goals. They have justified strict confidentiality through a conglomeration of abstract rationales that are difficult to test or contradict. n261 Empirical evidence concerning the effect of different types of exceptions on lawyer and client conduct would shed light on the strength of individual justifications. This in turn may encourage drafters to express the considerations that truly drive their rules.

D. Predicting the Effects of Different Exceptions

Consider a hypothetical jurisdiction that adopts a series of exceptions to a previously strict confidentiality rule. A client makes her attorney the following proposition: "If I double your fee, would you waive your right to disclose?" Because of the tempting nature of such offers, code drafters must decide in advance whether clients and attorneys should be allowed to change confidentiality contractually. n262 Empirical research can inform that decision.

It seems obvious that exceptions requiring disclosure of particular types of information should not be negotiable. Mandatory exceptions reflect a judgment that revelation of the information is in the public interest. Even if the attorney-client relationship is viewed as largely [*405] contractual, n263 some agreements must be deemed void. Waivers of mandatory disclosure provisions would violate public policy both in facilitating misconduct and in undermining society's right to receive information. n264

Discretionary exceptions raise the waiver issue more starkly. In other contexts, American law distinguishes between reaching discretionary decisions and coming to the same results by failing or refusing to exercise discretion. n265 Lawyer codes in general purport to place a premium on maintaining attorney independence. n266 In opting for discretionary rather than mandatory exceptions, code drafters probably would rely on the ability of lawyers to make informed case-by-case judgments as officers of the court. A major thrust of confidentiality exceptions is to make lawyers consider the ethics of their conduct and that of their clients. n267 Allowing lawyers to avoid confronting moral dilemmas by contract arguably would undermine the exceptions' goals.

Permitting lawyers to waive disclosure exceptions might well have perverse practical consequences. Under one hypothesis, all lawyers would sell their right to disclose. Absolute confidentiality rules then would clearly serve society better. They lead to the same results, while saving the transaction costs of bargaining.

One can, however, picture another world in which some lawyers cannot be bought cheaply. n268 Their clients would have an incentive to seek representation elsewhere. In particular, clients with much to hide would place a high premium on absolute confidentiality and would take care to insist upon a contractual silence agreement. n269 Confidentiality would thus become strictest when society most needs disclosure. The lawyers representing the secretive clients would be those with the least interest in acting [*406] morally. n270 Those lawyers are also less likely to persuade clients to act ethically by means other than disclosure. Under this scenario, code drafters might also prefer to avoid disclosure exceptions rather than adopting mutable exceptions.

There is, however, some reason to predict a third world in which the practical effects of allowing waiver might not be so significant. Under the current regime, some jurisdictions seem to allow lawyers to obtain advance client permission to disclose particular types of confidential information. n271 The fact that "moral" lawyers virtually never ask clients for the right to disclose illustrates the power of inertia. n272 The reality is that the rules set the standards for the profession. Attorneys and clients may not think about departing from professional norms. n273 Drafting even discretionary disclosure exceptions sends a message to clients and attorneys that disclosure in a particular context is the norm. Unless lawyers are prepared to advertise their willingness to act amorally as a means of producing business, they have little reason to mention to clients the subject of limiting their options. n274 They would be likely to accept the status quo in the codes and postpone any disclosure issue until the unlikely event a need to disclose arises.

In this hypothetical world, the impact of allowing waivers thus would come from clients who learn of the possibility of waiver independently, perhaps through frequent dealings with lawyers. Most probably, the degree of confidentiality would vary among classes of clients, in direct proportion [*407] to their sophistication. Such a distinction among clients is not justified by reference to the rationales for confidentiality or exceptions, but it is not unique. n275 Whether code drafters could countenance the distortion might well depend on how pronounced it is.

Empirical research can help gauge how lawyers would react to a rule allowing waiver. An investigation into lawyers' current practices -- their willingness or unwillingness to amend professional rules by agreement -- would inform the debate. Asking lawyers more directly whether they would sell a right to disclose particular information might not conclude the issue, but would provide drafters with useful and relevant information. The three hypotheses discussed above are all plausible and theoretically defensible. Without a factual foundation enabling some prediction of how lawyers and clients might act, code drafters analyzing the waiver issue are necessarily at sea.

E. Summary: Goals for Future Research

Empirical research can beguile. For example, learning that lawyers or clients favor a certain rule or exception tempts the researcher to support the provision. But as this Article's analysis of confidentiality's theoretical underpinnings makes clear, the matter is far more complex.

Strict confidentiality rules rest on assertions concerning the effects of rules on clients. Empirical research can support or cast doubt on the validity of those assertions, or may reveal independent considerations militating against reliance on confidentiality's justifications. Substantiating lawyers' and clients' normative views of confidentiality may illustrate how rules work and whether exceptions would create distrust of lawyers. But insofar as the research merely polls viewpoints, it justifiably deserves little respect.

The Tompkins County study was a trial balloon. Its questionnaires were designed to identify mistakes so that larger scale investigations would not fall prey to them. The Tompkins County experience defines a focus for future research. Subsequent studies, as a result, can direct their inquiry to particular aspects of confidentiality rules and determine how, in practice, particular exceptions would work.

[*408] CONCLUSION

Labelling attorneys "officers of the court" is no mere accident. A system encompassing substantial secret dealings between lawyer and client relies on lawyers as a screen and advocate against antisocial client conduct. But in extreme situations calling for intervention, strict confidentiality rules often reduce lawyers' ability and desire to act.

Allowing lawyers some leeway to reveal client information would not destroy our legal order. Other well-developed legal systems circumscribe confidentiality more narrowly than we. n276 Their continued viability suggests that our system, too, would survive limited exceptions. n277

At root, whether to adopt exceptions is a balance of countervailing costs and benefits. But the balance inevitably depends on facts. Only by determining how well strict confidentiality serves its intended purposes, and how much exceptions would undermine those goals, can code drafters put the normative question into proper perspective.

This Article's theoretical analysis and the data it presents are a first step toward providing that perspective. The Article unashamedly raises [*409] more questions than it answers. It is intended as a bridge to further study, and to show that empirical analysis of confidentiality rules is possible. The Article's presentation should assuage the fears of professional responsibility scholars who have resisted empirical research and provoke them to accept a challenge they have too long ignored.

APPENDIX -- HYPOTHETICALS

1. A client tells a lawyer the location of a "missing child" or kidnapping victim. The client is not implicated in the person's disappearance, but does not want the lawyer to disclose the information because the client "doesn't want to get involved." The client will not accept the lawyer's assurance that the lawyer could act without naming the client. n278

2. An attorney obtains information from a client that would prove that another person falsely accused of a crime, is innocent. The attorney could reveal the information without implicating the client in the crime. The client refuses to disclose the information voluntarily. n279

3. In a civil suit involving a serious automobile accident, the plaintiff is examined by defendant's doctor. The doctor discovers a life threatening aneurism. Although the aneurism was probably caused by the accident, plaintiff's own physician has not discovered it. The condition is curable, but if defendant's lawyer does not reveal the danger -- against the client's wishes -- plaintif may die. n280

4. The general counsel to a firm that produces a metal alloy used in the manufacture of airplanes learns of a company study that suggests that in some high-altitude flight patterns the alloy might weaken and cause a plane to explode. The alloy does, however, meet the minimum safety standards set by the government. The lawyer urges the Board of Directors to recall the alloy or at a minimum to inform users of its potential danger. The Board decides that the study is too inconclusive to warrant action in [*410] light of the dire financial consequences of disclosure to the company. n281

5. From information given by and conversations with a client, an attorney becomes convinced that the client is mentally imbalanced, out of control, and will injure someone in the near future. The client has, however, not expressed any specific intention to commit a crime. n282

6. X negotiates directly with Y and agrees to buy Y's house. Y agrees to provide "owner financing," subject to the contingency that X supply certain information concerning X's ability to pay. The only role X's attorney is to play in the transaction is to write the final sale contract and preside over the closing. Right before the closing, X's attorney learns from X's accountant that the financing information X supplied to Y contained inaccurate information. If any fraud was committed, it occurred before the attorney became involved. n283

7. A client fortuitously receives an undeserved payment from the government (e.g., a duplicate welfare check or tax refund) and deposits it in a savings account. The client then contacts his/her attorney who advises the client to return the money. The client refuses. n284

8. A lawyer represents reputed organized crime figures in civil litigation. During a deposition of a key witness against them, the witness reverses his testimony and tells a story that is completely favorable to the lawyer's clients. When confronted by the lawyer, one client admits that "a friend had a talk with the witness" but denies any subornation of perjury. The lawyer has no way of knowing what really happened.

9. A tax lawyer learns, from working with the books of a client, that the client has received large sums from the Russian government. The client [*411] will only say that the money is income for services rendered. The lawyer knows that the client has access to classified government documents.

10. A lawyer for a political organization learns that the client is secretly a Nazi front.

11. There are rumors that a former Vice President who resigned after pleading nolo contendere to tax evasion charges plans to run for public office again. He publishes a book in which he asserts that he at all times protested his innocence of any wrongdoing to his attorney. The attorney remembers that, in fact, the Vice President admitted accepting graft. n285

12. President Nixon announces publicly that the Watergate tapes show he knew nothing about any potentially illegal activities. Nixon's lawyer listens to the tapes and urges Nixon to disclose its contents "in the national interest." Nixon refuses and tells the press that he will not disclose anything about them because of his "duty to protect the presidency."

13. A car manufacturer knows that the placement of the gas tank on its compact car will cause the vehicle to burst into flames in 5% of rear end collisions. Its accountants determine that moving the gas tank will cost more than the company would have to pay in tort liability for injuries caused by the car's design. The company's attorney concludes that the company would not violate any criminal statute by continuing to produce the car, but nevertheless urges recall and redesign of the "defect." The company declines. n286

14. A lawyer represents a minister who regularly appears and solicits contributions on television and collects millions of dollars of contributions annually. The lawyer learns that the client plans to embezzle and ultimately abscond with the money. The lawyer's functions do not include assisting the client in the wrongdoing. n287

15. A lawyer advises a client that she has not satisfied a two-year residence requirement for filing for divorce within the jurisdiction. He later learns that she has filed for divorce, using another attorney, and has falsely sworn that she has resided in the jurisdiction for two years. n288

FOOTNOTES:
n1. The study, including the hypotheticals about which it asked, is described in detail infra at text accompanying notes 136-231. Computerized results are on file in my office and at the Iowa Law Review. Throughout the text, I refer to the study as the "Tompkins County study." I refer to specific pages of results as follows: (a) survey of lawyers -- "T.C. Table L , at"; (b) survey of laypersons -- "T.C. Table J , at ." Where the Article quotes comments by responding lawyers or laypersons, I refer to the individual's survey by the coded identification number -- "T.C. ID No. L " or "T.C. ID No. J ."

n2. N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 4-101 (1978). By adding facts, one could of course change the scenario to one in which the New York confidentiality rule permits disclosure. For example, if the client is a participant in the kidnapping -- before or after the fact -- the client information becomes a nonconfidential statement of an intent to commit a future or ongoing crime. Id. DR 4-101(C)(3).

n3. T.C. Table L1, at 16.

n4. T.C. Table J1A, at 43.

n5. The ABA's two most recent model codes recommend fairly strict confidentiality. The Model Rules provide:

A lawyer may reveal such information to the extent the lawyer reasonably believes necessary:

(1) to prevent the client from committing a criminal act that the lawyer believes is likely to result in imminent death or substantial bodily harm; or

(2) to establish a claim or defense on behalf of the lawyer in a controversy between the lawyer and the client, to establish a defense to a criminal charge or civil claim against the lawyer based upon conduct in which the client was involved, or to respond to allegations in any proceeding concerning the lawyer's representation of the client.

MODEL RULES OF PROFESSIONAL CONDUCT Rule 1.6(b) (1987) [hereinafter MODEL RULES]. In addition, a lawyer must disclose "material fact[s] to a tribunal when . . . necessary to avoid assisting a criminal or fraudulent act by the client." Id. Rule 3.3(a)(2). Delaware, Louisiana, and Missouri have adopted these Model Rules intact. I-IV NATIONAL REPORTER ON LEGAL ETHICS AND PROFESSIONAL RESPONSIBILITY (D. Luban ed. 1983). Arizona, Arkansas, Connecticut, Indiana, and Wyoming have enacted the substance of the Model Rules, but have expanded Rule 1.6(b)(1) to cover additional crimes. Id.

The 1969 Code of Professional Responsibility provides:

A lawyer may reveal:

. . . .

(3) The intention of his client to commit a crime and the information necessary to prevent the crime.

(4) Confidences or secrets necessary to establish or collect his fee or to defend himself . . . against an accusation of wrongful conduct.

MODEL CODE OF PROFESSIONAL RESPONSIBILITY DR 4-101(C) (1969) [hereinafter CPR]. A lawyer may also disclose information to rectify a client's "fraud upon a person or tribunal" if he receives information "clearly establishing" the fraud and the information is not "protected as a privileged communication." Id. DR 7-102(B). Colorado, Iowa, Massachusetts, Tennessee, Vermont, and West Virginia maintain essentially unedited versions of the CPR. I-IV NATIONAL REPORTER ON LEGAL ETHICS AND PROFESSIONAL RESPONSIBILITY (D. Luban ed. 1983). Alaska, Georgia, Hawaii, Kansas, Michigan, Minnesota, Ohio, Oklahoma, Pennsylvania, Rhode Island, South Carolina, and Texas follow the basic format of the CPR, but allow somewhat more disclosure when a client intends to commit a fraud. Id.

Other states have diverged substantially from the spirit of the model codes. Among the most protective of confidences, New York, North Carolina, and Washington allow disclosure to prevent crime, but require confidentiality to be maintained for all noncriminal fraud. Id. At the other extreme, New Jersey and Wisconsin require disclosure to prevent crime or fraud likely to cause substantial harm and allow disclosure to rectify a crime or fraud in which the lawyer's services were used. Id.

n6. See infra text accompanying notes 85-91.

n7. See, e.g., Fried, Too High a Price for Truth: The Exception to the Attorney-Client Privilege for Contemplated Crimes and Frauds, 64 N.C.L. REV. 443, 490-91 & n.270 (1986); Morgan, Foreword to MODEL CODE OF EVIDENCE 27 (1942); Rhode, Ethical Perspectives on Legal Practice, 37 STAN. L. REV. 589, 613-14 (1985) & authorities cited at nn. 88-89; Thurman, Limits to the Adversary System: Interests that Outweigh Confidentiality, 5 J. LEG. PROF. 5, 9 (1980); Note, The Attorney-Client Privelege: Fixed Rules, Balancing, and Constitutional Entitlement, 91 HARV. L. REV. 464, 470 (1977).

n8. Other explanations are possible. Professional responsibility scholars tend to distrust empirical research. Some believe establishing ethical propositions factually is an impossible task. Others fear their own inability to interpret data intelligently, even if it becomes available. See AALS Workshop on Professional Responsibility, Keeping Guilty Secrets (March 5, 1988) (tape recording, tapes 9-10) (expressing such concerns).

n9. But see Marks & Cathcart, Discipline Within the Legal Profession: Is it Self Regulation?, 1974 U. ILL. L.F. 193 (noting and explaining lack of enforcement of many code violations).

n10. For example, in rejecting proposed exceptions to the Model Rules' strict confidentiality provisions, the ABA appeared to kowtow to the selfish interests of the corporate bar. See, e.g., Abel, Why Does the ABA Promulgate Ethical Rules?, 59 TEX. L. REV. 639, 655-56 (1981); Rhode, Why the ABA Bothers: A Functional Perspective on Professional Codes, 59 TEX. L. REV. 689, 691-92 (1981).

n11. See In re Primus, 436 U.S. 412, 434 (1978) (association and political speech); Bates v. State Bar of Ariz., 433 U.S. 350, 383 (1977) (commercial speech).

n12. See, e.g., Shapero v. Kentucky Bar Ass'n, 108 S. Ct. 1916, 1923 (1988) (solicitation); Supreme Court of N.H. v. Piper, 470 U.S. 274, 282-83 (1985) (pro hac vice rule); In re R.M.J., 455 U.S. 191, 199 (1982) (advertising); Hirschkop v. Snead, 594 F.2d 356, 365-66 (4th Cir. 1979) (trial publicity).

n13. In Zacharias, Rethinking Confidentiality II: Is Confidentiality Constitutional?, 75 IOWA L. REV. (1989) (forthcoming), I consider whether strict confidentiality rules violate lawyers' first amendment rights. Whatever legal analysis courts adopt to assess the constitutional issue, they will have to focus on the importance of the state interests underlying confidentiality and the degree to which an inflexible rule is necessary to serve those interests.

n14. Cf. Burt, Conflict and Trust Between Attorney and Client, 69 GEO. L.J. 1015, 1018 (1981) ("The proponents and opponents of the [proposed Model Rules on confidentiality were] arguing past one another -- the former preferring 'public protection,' the latter preferring 'client protection,' and neither [were] able to demonstrate the social value of preferring one over the other." (footnotes omitted)).

n15. Arguably, nonenforcement incorporates de facto exceptions into the strict confidentiality rules, thereby changing the prevailing law without forethought and debate.

n16. Cf. Leubsdorf, Three Models of Professional Reform, 67 CORNELL L. REV. 1021, 1025 (1982) ("[T]he traditional emphasis on the attorney-client relationship is too narrow. Providing a client with . . . devoted counsel increases the likelihood that the two will inflict damage on opposing and unrepresented parties.").

n17. See supra note 5.

n18. See infra text accompanying notes 22-113. This Article distinguishes attorney-client confidentiality from the narrower concept of attorney-client privilege and confines itself to nonevidentiary confidentiality rules. Compare 8 J. WIGMORE, EVIDENCE § 2292, at 554 (J. McNaughton rev. 1961) (privilege applies only to attempts in litigation to compel lawyers to disclose client communications) and United States v. United Shoe Mach. Corp., 89 F. Supp. 357, 358-59 (D. Mass. 1950) with J. BURKOFF, CRIMINAL DEFENSE ETHICS § 6.4(a)(1), at 6-32 (1986) and C. WOLFRAM, MODERN LEGAL ETHICS § 6.7.3, at 299 (1986) (discussing reasons why lawyer codes forbid disclosure of confidential information in any setting); see also L. PATTERSON, LEGAL ETHICS: THE LAW OF PROFESSIONAL RESPONSIBILITY § 6.03, at 205 (1984) (noting that Canon 4 of CPR protects "all information given by a client to his attorney whether or not strictly confidential in nature"). Unlike confidentiality, the privilege has judicial or legislative roots. See 9 W. HOLDSWORTH, A HISTORY OF ENGLISH LAW 198-203 (1926); 8 J. WIGMORE, supra, §§ 2290, 2291 and authorities therein; see also L. PATTERSON, supra, § 1.02 (describing United States ethical standards dating back to 1836); C. WOLFRAM, supra, § 2.3, at 36, § 2.6.2, at 55-56, § 2.6.3, at 56-58 (source of 1908 Canons and subsequent codes were lawyer-drafted proposals, rubber-stamped by legislatures or governing judicial panels); Patterson, Legal Ethics and the Lawyer's Duty of Loyalty, 29 EMORY L.J. 909, 912-13 (1980) (same). The privilege is well defined by case law. See, e.g., C. MCCORMICK, EVIDENCE § 87, at 204-08 (E. Cleary 3d ed. 1984); 8 J. WIGMORE, supra, § 2286, at 530-31, § 2290, at 542-45. Although the justifications for the privilege mirror those advanced in favor of confidentiality, see generally Gardner, A Reevaluation of the Attorney-Client Privilege, 8 VILL. L. REV. 279, 289-339 (1963) (discussing roots of privilege in furthering direct communication); Hazard, An Historical Perspective on the Attorney-Client Privilege, 66 CALIF. L. REV. 1061, 1069-91 (1978) (same), additional systemic considerations may contribute to the privilege's current formulation. Considerations that may be important to the privilege, but not to confidentiality, include judicial economy, respect for precedent, and the separate societal interest in obtaining evidence necessary to reach fair judicial decisions.

n19. See infra text accompanying notes 115-231.

n20. This Article's point is simply to highlight heretofore unrecognized costs and effects of strict confidentiality. The Article thus neither advocates particular disclosure exceptions nor takes a position on whether rule changes should be discretionary (i.e., "lawyers may disclose") or mandatory (i.e., "lawyers must disclose"). Each format will have different effects on clients. See, e.g., infra text accompanying notes 262-75. A mandatory disclosure exception may, for example, enhance society's view of lawyer integrity while doing nothing to encourage lawyers to think independently in moral terms. Conversely, a discretionary rule may force some lawyers to engage in moral self-examination, but allows others to avoid doing "the right thing" in favor of self-serving practical concerns. The precise context for which code-drafters frame the rules should determine the preferable approach.

n21. See infra text accompanying notes 232-75.

n22. Compare Frankel, The Search for Truth Continued: More Disclosure, Less Privilege, 54 U. COLO. L. REV. 51, 53 (1982) [hereinafter Frankel, Search for Truth Continued] (the legal system would improve if lawyers were required to disclose all information favorable to a client's adversary) and Frankel, The Search for Truth: An Umpireal View, 123 U. PA. L. REV. 1031, 1052 (1975) [hereinafter Frankel, Umpireal View] (same) with Alschuler, The Search for Truth Continued, the Privilege Retained: A Response to Judge Frankel, 54 U. COLO. L. REV. 67 (1982) [hereinafter Alschuler, The Search for Truth] (questioning extent to which modification of confidentiality and attorney-client privilege would remedy excesses of adversarial system), Alschuler, The Preservation of a Client's Confidences: One Value Among Many or a Categorical Imperative, 52 U. COLO. L. REV. 349, 350 (1981) [hereinafter Alschuler, The Preservation of a Client's Confidences] (Judge Frankel's framing of problem is misleading and oversimplified) and Uviller, The Advocate, the Truth and Judicial Hackles: A Reaction to Judge Frankel's Idea, 123 U. PA. L. REV. 1067, 1067 (1975) (the adversarial system is "one of the better methods" for reconstructing truth).

n23. All of the existing codes encourage lawyers to discuss ethical dilemmas with their clients and to encourage clients to authorize appropriate, ethical conduct. In many situations, lawyers can exert pressure on their clients by withdrawing or threatening to withdraw. Even when disclosure of a client secret seems the only suitable alternative, loose language often enables lawyers to find ways to circumvent the proscriptions of confidentiality rules. Arguably, even strict codes like the Model Rules leave room for ethical behavior that would satisfy society's interests in disclosure.

Nevertheless, it begs the question to argue that actual dilemma situations arise infrequently or that lawyers can circumvent even strict rules when these situations do arise. In the typical situation, professional codes necessarily affect lawyer attitudes and conduct. When strict rules forbid disclosure, at least some lawyers feel bound to silence.

n24. Much of the literature in support of broad confidentiality concerns the criminal context. See generally M. FREEDMAN, LAWYERS' ETHICS IN AN ADVERSARY SYSTEM (1975); Freedman, Are the Model Rules Unconstitutional?, 35 U. MIAMI L. REV. 685, 690 (1981); Wolfram, Client Perjury, 50 S. CAL. L. REV. 809 (1977). But see Pepper, The Lawyer's Amoral Ethical Role: A Defense, A Problem, and Some Possibilities, 1986 AM. B. FOUND. RES. J. 613, 621-24 (noting the civil-criminal distinction and proposing a model to justify lawyers' "amorality" in all contexts). Robert Gordon points out that outside the criminal arena the scholarly consensus approving extreme adversarial representation evaporates. Gordon, The Independence of Lawyers, 68 B.U.L. REV. 1, 11 (1988).

n25. See, e.g., Flemming, Client Games: Defense Attorney Perspectives on Their Relations with Criminal Clients, 1986 AM. B. FOUND. RES. J. 253 (empirical study concluding that criminal defendants distrust appointed attorneys more than civil clients distrust their retained attorneys); see also A. GOLDMAN, THE MORAL FOUNDATIONS OF PROFESSIONAL ETHICS 117-20 (1980) (maximum protection of possibly innocent criminal defendants not necessary in civil context since it is not individual against state); Gillers, Can a Good Lawyer Be a Bad Person?, 84 MICH. L. REV. 1011, 1024-25 (1986) (greater social need to protect criminal lawyers since they play more essential role in adversary system); Luban, The Adversary System Excuse, in D. LUBAN, THE GOOD LAWYER 83, 91-93, 115-18 (1983) (more protection for criminal lawyers necessary since goal is protection of accused from state rather than assignment of remedies); Schwartz, The Zeal of the Civil Advocate, in D. LUBAN, supra, at 150, 155-56, 160 (justifications for criminal lawyers' confidentiality behavior not present in civil context).

n26. The sixth amendment assures criminal defendants effective counsel, Kimmelman v. Morrison, 477 U.S. 365, 377-78 (1986), and thus perhaps some degree of confidentiality. See Brazil, Unanticipated Client Perjury and the Collision of Rules of Ethics, Evidence, and Constitutional Law, 44 MO. L. REV. 601, 623-24 (1979); Callan & David, Professional Responsibility and the Duty of Confidentiality: Disclosure of Client Misconduct in an Adversary System, 29 RUTGERS L. REV. 332, 367 (1976) and authorities cited therein; Seidelson, The Attorney-Client Privilege and the Client's Constitutional Rights, 6 HOFSTRA L. REV. 693, 707-09 (1978); Subin, The Lawyer as Superego: Disclosure of Client Confidences to Prevent Harm, 70 IOWA L. REV. 1091, 1127-34 (1985) and authorities cited therein; Wolfram, supra note 24, at 840-42; cf. United States v. Henry, 447 U.S. 264, 295 (1980) (Rehnquist, J., dissenting) ("the Sixth Amendment, of course, protects the confidentiality of communications between the accused and his attorney"). Other constitutional provisions also help justify a greater concern for confidentiality in the criminal context than in the civil. The privilege against self-incrimination, for example, prohibits government actions that require criminal defendants or their attorneys to provide evidence. See, e.g., Fisher v. United States, 425 U.S. 391, 402, 403-05 (1976) (lawyer may be required to disclose documents that could be subpoenaed from client, but may not disclose documents prosecution could not obtain if in client's hands); United States v. Judson, 322 F.2d 460, 467 (9th Cir. 1963) (same); see also Subin, supra, at 1091, 1119-27 and authorities cited therein. Commentators advocating strict confidentiality have also relied on the right to privacy, see, e.g., Note, supra note 7, at 483-84, and a combination of the fifth and sixth amendments. See id. at 485-86; Note, Functional Overlap Between the Lawyer and Other Professionals: Its Implications for the Privileged Communications Doctrine, 71 YALE L.J. 1226, 1236-37 (1962).

n27. My point is not that confidentiality in the criminal context should be absolute, only that it presents a different issue. See D. LUBAN, LAWYERS AND JUSTICE 202-05 (1988) (discussing arguments that have unique merit in criminal cases); Schwartz, supra note 25, at 155-56 (discussing reasons why morality of criminal defense different from that in civil representation); Wasserstrom, Lawyers as Professionals: Some Moral Issues, 5 HUM. RTS. 1, 12 (1975) ("the amoral behavior of the criminal defense lawyer is justifiable . . . [because] of the special needs of the accused"); cf. Eshete, Does a Lawyer's Character Matter?, in D. LUBAN, supra note 25, at 270, 272-74 (pure adversarial model more appropriate in criminal context). For even the constitutional defense of confidentiality has limits. For example, courts have exempted attorney disclosure of some types of information from the ambit of the privilege against self-incrimination. The government may be able to compel an attorney to disclose particular documents which, if in the hands of the client, would be subject to fifth amendment protection. See, e.g., Fisher v. United States, 425 U.S. 391, 400-01 (1976); United States v. White, 477 F.2d 757, 763, aff'd, 487 F.2d 1335 (5th Cir. 1973) (en banc), cert. denied, 419 U.S. 872 (1974). Similarly, a lawyer may be required to turn over incriminating physical evidence that the client could hide under the privilege. See, e.g., In re Ryder, 263 F. Supp. 360, 365 (E.D. Va.), aff'd, 381 F.2d 713 (4th Cir. 1967); Morrell v. State, 575 P.2d 1200, 1211 (Alaska 1978); State ex rel. Sowers v. Olwell, 64 Wash. 2d 828, 833-34, 394 P.2d 681, 684 (1964). See generally J. BURKOFF, supra note 18, § 6.5(h) (discussing the conflicting authorities); Freedman, Professional Responsibility of the Criminal Defense Lawyer: The Three Hardest Questions, 64 MICH. L. REV. 1469 (1966); Griffiths, Ideology in Criminal Procedure or a Third "Model" of the Criminal Process, 79 YALE L.J. 359, 404-09 (1970) (debating this issue); Noonan, The Purposes of Advocacy and the Limits of Confidentiality, 64 MICH. L. REV. 1485 (1966). Some dispute remains as to what an attorney may disclose if a client threatens to commit perjury, but it is now clear that the attorney need not rest idle. See Nix v. Whiteside, 475 U.S. 157, 176 (1986).

n28. See, e.g., ABA COMM'N ON EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT 82 (Proposed Alternative Final Draft 1981); C. WOLFRAM, supra note 18, at § 6.1.3; Morgan, supra note 7, at 25-26.

n29. See, e.g., Freedman, supra note 27, at 1473; Freedman, supra note 24, at 689-94 (Model Rules do not sufficiently protect confidentiality).

n30. See, e.g., Natta v. Hogan, 392 F.2d 686, 691 (10th Cir. 1968); United States v. Grand Jury Investigation, 401 F. Supp. 361, 369 (W.D. Pa. 1975); see also ABA COMM'N ON EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT 38 (Proposed Final Draft 1981) and authorities cited at 42.

n31. Arguably, a lawyer needs to learn all relevant facts in order to be effective. See, e.g., ABA COMM'N ON EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT 42 (Proposed Final Draft 1981); Gardner, supra note 18, at 318; Hazard, supra note 18, at 1061. Absent confidentiality a lawyer might, for fear of having to disclose, make himself less effective by avoiding damaging information by his clients. See Brazil, supra note 26, at 641-42; Gardner, supra note 18, at 319. Most justifications for confidentiality rules therefore rely on the three-step syllogism discussed supra text accompanying note 28: (1) it is useful for society to make legal advice available; (2) in order to give legal advice, lawyers need full information; (3) to obtain full disclosure, confidentiality must be guaranteed. See C. WOLFRAM, supra note 18, § 6.1.3.

n32. The adversary system is based on the theory that if two fully equipped, fully informed attorneys do battle in the courtroom, the truth will win out. See S. BOK, LYING: MORAL CHOICE IN PUBLIC AND PRIVATE LIFE 161 (1978); S. LANDSMAN, THE ADVERSARY SYSTEM 2 (1984); Lorne, The Corporate and Securities Adviser, the Public Interest, and Professional Ethics, 76 MICH. L. REV. 425, 473 (1978); Schwartz, supra note 25, at 153-55.

n33. Confidentiality serves the notion that every client is entitled to one ally with undivided loyalty. See, e.g., S. BOK, supra note 32, at 159-60; M. FREEDMAN, supra note 24, at 4; Alschuler, The Search for Truth, supra note 22, at 73. It underscores the concept that the lawyer is the client's agent -- his fiduciary. See ABA COMM'N OF EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT 172-73 (Proposed Final Draft 1981) and authorities cited therein. Looking at it from a different perspective, clients might distrust lawyers when they maintain a right to disclose embarrassing or harmful information. See Brazil, supra note 26, at 642. It seems unsavory for lawyers first to induce disclosure and then make the information public. See Alschuler, The Preservation of a Client's Confidences, supra note 22, at 351.

n34. See Alschuler, The Search for Truth, supra note 22, at 73; Pepper, supra note 24, at 630-32. Sissela Bok notes that a system shows respect for the autonomy of individuals when it allows them to have and to share secrets. S. BOK, SECRETS 120-21 (1982); see also Radin, The Privilege of Confidential Communication Between Lawyer and Client, 16 CALIF. L. REV. 487, 487-93 (1928) (tracing place of Roman notions of "mutual fidelity" and "gentleman's honor" in confidentiality rules). Conversely, allowing attorney disclosures might enable lawyers to blackmail clients into following their advice. It thus would disrespect the client's ability to make personal decisions and elevate the lawyer's role from the client's agent to principal. See Lorne, supra note 32, at 473; see also ABA COMM'N ON EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT 42 (May 30, 1981 draft) (discussing role of lawyer as the client's agent); cf. Rhode, supra note 7, at 605 (describing and questioning argument that the code in general enhances client "dignity").

n35. See American Bar Association, Statement of Policy Regarding Lawyers' Responses to Auditors' Requests for Information, 31 BUS. LAW. 1709-10 (1976). Similarly, full information helps the lawyer avoid spurious lawsuits. See C. WOLFRAM, supra note 18, § 6.1.4, at 244. In general, if confidentiality encourages the use of counsel, it also may promote compliance with the law. See Lorne, supra note 32, at 473.

n36. Arguably, if lawyers had discretion to disclose such information, unethical lawyers would profit by selling a guarantee of silence. Strict confidentiality rules have the effect of spreading the benefits throughout the guild.

n37. See Wolfram, supra note 24, at 809, 839-40. There are, of course, two faces to this shield. Sometimes, preventing harm to lawyers' reputations due to affiliation with notorious clients may benefit society, for we often want lawyers to take on unpopular causes. Fiction and nonfiction libraries alike are filled with stories of lawyers who have risen above popular opinion and rescued the innocent. See, e.g., R. BARTELS, BENEFIT OF LAW (1988) (relating author's vindication of individual wrongly imprisoned for notorious, sexually related murder of child); H. LEE, TO KILL A MOCKINGBIRD (1960) (fictional account of attorney representing falsely accused black man in rural South); A. WEINBERG, ATTORNEY FOR THE DAMNED: CLARENCE DARROW IN HIS OWN WORDS (1957) (describing unpopular causes represented by Clarence Darrow). Nevertheless, attorneys who hide behind confidentiality in order to represent unpopular clients often do not do so for selfless reasons.

n38. See, e.g., MODEL RULES Rule 1.6(b)(2); CPR DR 4-101(C)(4).

n39. America, to a far greater extent than other societies, accepts the hired gun paradigm. See, e.g., D. RUESCHEMEYER, LAWYERS AND THEIR SOCIETY 128 (1973) (noting "emphasis in the German writings on legal ethics on the duty to maintain one's independence from the client"); Luban, The Sources of Legal Ethics: A German American Comparison of Lawyers' Professional Duties, 48 RABELS ZEITSCHRIFT FUR AUSLANDISCHES UND INTERNATIONALES PRIVATRECHT 245, 267 (1984) (German proceedings require "much greater independence of the lawyer from his client's desires than in America"); Taruffo, The Lawyer's Role and the Model of Civil Process, 16 ISRAEL L. REV. 5, 13-19 (1981) (discussing, inter alia, Austrian, Italian, and Soviet systems and concluding that civil law systems do limit attorneys' freedom to act on behalf of clients, but perhaps not to as great an extent as some commentators presume). The hired gun rationale is, in part, the justification used to explain why lawyers are forbidden to reveal client secrets. See, e.g., M. FREEDMAN, supra note 24, at 1-8; D. LUBAN, supra note 25, at 10-12; Donagan, Justifying Legal Practice in the Adversary System, in D. LUBAN, supra note 25, at 123, 139-45.

n40. Professor Patterson suggests that such considerations initially provoked the profession to formulate and accept the modern view of confidentiality. Patterson, supra note 18, at 915-16, 951-55. The hired gun approach, of course, also may help the lawyer avoid self-doubt as to whether representing a disreputable client or representing the client in a particular way is morally justifiable. Cf. Kennedy v. Kennedy, 20 Mass. App. Ct. 559, 564, 481 N.E.2d 1172, 1176 (1985) (Brown, J., concurring) ("There must be limits on what a lawyer will do for a client. Lawyers are not hired guns. Our profession is a higher calling, and we should be about loftier business."). More generally, it assists the bar in fending off outside regulation that might evaluate the professional conduct of individual attorneys. Under the hired gun definition, society views the attorney merely as an extension of the client. Regulation of the client thus seems to suffice to satisfy society's interest in any particular case.

n41. See Morgan, The Evolving Concept of Professional Responsibility, 90 HARV. L. REV. 702, 707-12 (1977) (discussing "legal" functions nonlawyers might perform).

n42. I would not suggest that code drafters write rules with a view to their own finances. Such considerations are, however, likely to weigh heavily on the minds of the drafters' constituents -- the lawyers the rules affect. The ABA membership, for example, recently watered down several proposed model rules that adversely affected lawyers' personal interests -- including some relating to confidentiality. See, e.g., Landesman, Confidentiality and the Lawyer-Client Relationship, 1980 UTAH L. REV. 765, 766-69. See generally Gillers, supra note 25, at 1017 ("the lawyers' near monopoly on the scope of their professional duty has often resulted in narrow and self-interested resolutions"); Gillers, What We Talked About when We Talked About Ethics: A Critical View of the Model Rules, 46 OHIO ST. L.J. 243, 245, 256 (1985) (Model Rules are internally inconsistent to Bar's benefit); Morgan, supra note 41, at 739 (Code effectively ends attorney's obligation of confidentiality when it becomes uncomfortable for attorney); Rhode, supra note 7, at 600-01; Rhode, supra note 10, at 711.

n43. Sissela Bok has pointed out the general dangers of letting lawyers set their own standards. S. BOK, supra note 32, at 117; see also Rhode, supra note 7, at 590, 616, 641 (discussing "parochial" orientation of lawyer drafters of the codes). Whether the benefits of confidentiality outweigh its costs is a public issue on which laypersons and lawyers may differ. See Note, supra note 26, at 1232 (lawyers have a far different perception of confidentiality than laypersons); see also infra text accompanying notes 189-95 (Tompkins County study suggests lawyers misperceive how citizens would react to loosening of confidentiality rules). If drafters hide the thrust of their rules, the chances for public participation and comment decrease. Not only does the quality of the draft suffer, but a real danger of provisions affirmatively harmful to the public interest develops.

n44. Professor Elliston writes that legal professionalism would be enhanced by "recogniz[ing] the centrality of ethics to the practice of law and making a concerted effort to develop the conscience of the lawyer." Elliston, Ethics, Professionalism and the Practice of Law, 16 LOY. U. CHI. L.J. 529, 546 (1985). In suggesting that disclosure of client secrets sometimes may be appropriate, this Article, like Elliston, recognizes that such a change may require new "limits . . . on the client's use of lawyers to achieve immoral ends." Id.

n45. In reviewing a draft of this Article, John Leubsdorf recognized another possible justification for confidentiality rules. Ordinarily, private parties entering a commercial relationship have broad leeway in defining the degree to which trade secrets may be disclosed. Arguably, in the absence of rules, sophisticated clients would negotiate for confidentiality. Attorney-client confidentiality rules obviate the bargaining costs and, in the tradition of consumer protection, protect less educated clients who would not think of insisting upon secrecy.

I do not dwell on this rationale here. For while it may support the general notion of confidentiality, it does not necessarily justify the type of strict rule that is this Article's focus. Even contractual secrecy agreements may be subject to the superior rights of third parties. Indeed, agency law seems to provide for confidentiality in precisely the way Professor Leubsdorf envisions, but creates an exception for the protection of third parties. See infra text accompanying notes 46-49.

n46. RESTATEMENT (SECOND) OF AGENCY § 395 (1957); see also RESTATEMENT (SECOND) OF TRUSTS § 2 comment b (1957) ("A confidential relation exists between two persons when one has gained the confidence of the other and purports to act or advise with the other's interest in mind. A confidential relation may exist although there is no fiduciary relation. . . .").

n47. RESTATEMENT (SECOND) OF AGENCY § 396(c) (1957); see also RESTATEMENT (SECOND) OF TRUSTS § 44 comment c, § 45 comment c (1957); RESTATEMENT OF RESTITUTION § 200 comment a (1936).

n48. RESTATEMENT (SECOND) OF AGENCY § 395 comment f (1957); see also Willig v. Gold, 75 Cal. App. 2d 809, 814, 171 P.2d 754, 757 (1946) (agent is under no duty to keep confidential principal's dishonest acts); Hale v. Mason, 160 N.Y. 561, 567, 55 N.E. 202, 204 (1899) (trustee may disclose principal's misrepresentation to affected third party); cf. Pearson v. Dodd, 410 F.2d 701, 705 n.19, cert. denied, 395 U.S. 947 (1969) (noting argument that congressional staff members may have been privileged to disclose office documents revealing employer's wrongdoing); RESTATEMENT OF TORTS § 757 comment d (1939) (noting privilege to reveal trade secrets "to protect [a conflicting] interest of the actor").

n49. Many jurisdictions, for example, require doctors to report gunshot wounds suffered by patients and evidence that children may have been subjected to physical abuse. See, e.g., ILL. ANN. STAT. ch. 38, § 206-3.2 (West 1973) (medical professional must notify local law enforcement agency of treatment of injury resulting from discharge of firearm); N.Y. PENAL LAW § 265.25 (McKinney 1979) (same); ILL. ANN. STAT. tit. 23, § 2054 (West 1988) (medical professional must report reasonable suspicion of child abuse); N.Y. SOC. SERV. LAW § 413 (McKinney 1973) (same); cf. Chesky v. United States, No. 85-0487-B (D. Me. Mar. 1, 1988), (LEXIS, Genfed library, Courts file) (under Maine exception allowing disclosure to avoid harm to others, lawyer may reveal client's proclivity to abuse his son).

n50. See supra note 5 and accompanying text.

n51. Cf. RULES REGULATING THE FLORIDA BAR, Rule 4-1.6(b)(2) (requiring lawyers to reveal information "to prevent a death or substantial bodily harm to another"); MAINE BAR RULES, Rule 3.6 (lawyer may disclose to avoid subjecting others to risk of harm).

n52. Some states, however, including Connecticut, Maryland, Nevada, New Jersey, South Dakota, and Wisconsin, do allow attorneys to disclose in order to "rectify" frauds or crimes in which their services have been used. I-IV NATIONAL REPORTER ON LEGAL ETHICS AND PROFESSIONAL RESPONSIBILITY (D. Luban ed. 1983).

n53. Numerous commentators have advocated the need for a "moral dialogue" between attorney and client. See, e.g., Pepper, supra note 24, at 630-32; Shaffer, The Practice of Law as Moral Discourse, 55 NOTRE DAME L. REV. 231, 241-42 (1979). Others have suggested the creation of an "informed consent" doctrine for the legal profession, based in part upon the theory that the resulting attorney-client discussions will produce more ethical behavior. See Martyn, Informed Consent in the Practice of Law, 48 GEO. WASH. L. REV. 307, 316-18 (1980); Spiegel, Lawyering and Client Decisionmaking: Informed Consent and the Legal Profession, 128 U. PA. L. REV. 41, 118 (1979).

n54. In a few of the hypotheticals, the legitimacy of disclosure depends on the particular "strict" rule that applies. Under the Model Rules, for example, a lawyer with sufficient reason to fear his client had suborned perjury might be able to use confidences to prevent a fraud on the tribunal. See MODEL RULES Rule 3.3(a)(2). In New York and Washington, confidentiality must be preserved. See N.Y.S. Bar Association, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 7-102(B)(1) (1978); WASHINGTON RULES OF PROFESSIONAL CONDUCT Rule 3.3(c).

n55. Attorney-client privilege and confidentiality had altogether different roots. Originally, litigants were not permitted to testify. To avoid circumvention of this prohibition through the testimony of lawyers (i.e., relaying the client's information), nineteenth-century courts developed privilege rules. See Hazard, supra note 18, at 1082-83; Morgan, supra note 7, at 24; Note, supra note 7, at 466 and authorities cited therein. Subsequently, the rules evolved into a personal privilege belonging to the attorney. They essentially became a "gentleman's code of honor not to reveal secrets. See Gardner, supra note 18, at 289; Morgan, supra note 7, at 24-25.

Confidentiality became a client's right only in relatively modern times, based on general notions of its importance to the legal system. See Patterson, supra note 18, at 914-15, 941 (confidentiality "as an ethical duty . . . makes its first appearance as such in the [1887] Alabama Code of Ethics"). "Standards" for lawyer conduct previously had constrained lawyers from disclosing confidences. But for the most part, these did not entitle clients to lawyer silence. 2 D. HOFFMAN, A COURSE OF LEGAL STUDY 752 (2d ed. 1836) (stating "Resolutions in Regard to Professional Deportment"). Professor Patterson suggests that only with the adoption of the CPR in 1969 did the ethical rules enable "the lawyer to reject [personal] responsibility for non-disclosure." Patterson, supra note 18, at 946.

n56. Cf. Rhode, supra note 7, at 613 ("Defenders of broad confidentiality protections almost invariably assume what is to be proven, namely that any disclosure responsibilities would dismember lawyer-client relationships.") and authorities cited at nn. 88-89.

n57. Morgan, supra note 7, at 25.

n58. As one commentator has noted:

Clients seek out attorneys and will continue to do so largely because there is no ready substitute for legal advice and, perhaps, because they realize that the costs of withholding information from their legal representative are likely to outweigh the consequences which might result from any compelled disclosure of confidential communications.

Note, supra note 7, at 470-71. See generally Gardner, supra note 18, at 318; Subin, supra note 26, at 1164; Thurman, supra note 7, at 9. Professor Shavell has framed a more sophisticated economic model which shows that the degree to which confidentiality encourages clients to seek legal assistance depends on a variety of factors, including the likely effect of the advice and the client's state of mind before seeking advice. Shavell, Legal Advice About Contemplated Acts: The Decision to Obtain Advice, Its Social Desirability, and Protection of Confidentiality, 17 J. LEGAL STUD. 123, 129-30, 133-34, 136, 137-39 (1988).

n59. But see Simon, Ethical Discretion in Lawyering, 101 HARV. L. REV. 1083, 1142 (1988) ("People would have ample incentives to disclose adverse information to counsel even without confidentiality safeguards because they are honest and law-abiding, because they cannot make reliable judgments about when it is in their interests to withhold, or because in many business contexts they risk liability by failing to seek good legal advice."). Of, course, clients may withold information for a host of reasons unrelated to the existence of confidentiality, including mistrust of the attorney, embarrassment about the information, and his or her own view of what the attorney should know. See K. MANN, DEFENDING WHITE COLLAR CRIME 40-52 (1985).

n60. Rhode, supra note 7, at 614. The fact that clients continue to speak with their attorneys despite this confusion about confidentiality suggests that the status of the rules is not key to the decision to confide. See Schwartz, The Professionalism and Accountability of Lawyers, 66 CALIF. L. REV. 669, 684 (1978) (doubting that clients would be willing to risk liability due to insufficient disclosure).

n61. See, e.g., CPR DR 4-101(C)(3) and MODEL RULES Rule 1.6(b)(1) (future crimes); CPR DR 4-101(C)(4) and MODEL RULES Rule 1.6(b)(2) (lawyer self-protection); CPR DR 7-102 and MODEL RULES Rule 3.3 (fraud to tribunal and "candor" to tribunal).

n62. See, e.g., Fried, supra note 7, at 498 (trend towards more crime-fraud exceptions should be reversed); Levine, Self-Interest or Self-Defense: Lawyer Disregard of the Attorney-Client Privilege for Profit and Protection, 5 HOFSTRA L. REV. 783, 786 (1977) (noting difficulty in reconciling the expansion of the "to protect the attorney's interest" exception with the bar's general defense of the attorney-client privilege); Wolfram, supra note 24, at 837 & n.106 ("Conjuring up real-life hypothetical situations in which [the disclosure requirement for fraud on the tribunal applies] has become something of a law school parlor game.").

n63. Note, supra note 7, at 471. The CPR's "future crimes" exception is an example of a vague confidentiality loophole. See Fried, supra note 7, at 444 (noting difficulty of distinguishing between past and future crimes); see also id. at 461-89 (illustrating reasons why future crime exception is difficult to pinpoint). Arguments have long raged within the profession over what constitutes a "future" crime and what must be disclosed. See, e.g., M. FREEDMAN, supra note 24, at 27-42, 51-77; Kramer, Clients' Frauds and Their Lawyers' Obligations: A Study in Professional Irresponsibility, 67 GEO. L.J. 991, 994 (1979); Lawry, Lying, Confidentiality, and the Adversary System of Justice, 1977 UTAH L. REV. 653, 689. Except for clear cases, clients are unlikely to understand what lawyers themselves cannot.

To the extent confidentiality rules incorporate attorney-client privilege rules, more confusion may develop. For example, advice that a court concludes is "business" rather than "legal" advice may be revealed in open court. See United States v. United Shoe Mach. Corp., 89 F. Supp. 357, 359-60 (D. Mass. 1950) (making business/legal distinction); see also Note, supra note 7, at 471 (business/legal distinction is discretionary judgment "informed by a judge's ad hoc notion of what is reasonable in a given case"); cf. Underwater Storage, Inc. v. United States Rubber Co., 314 F. Supp. 546, 548 (D.D.C. 1970) (attorney not giving legal advice because he was "merely performing tasks which could as easily have been done by non-lawyers"). The client is unlikely to know in advance specifically what information fits within the two categories.

n64. See Thurman, supra note 7, at 9; cf. Gardner, supra note 18, at 318 (because litigants need lawyers, it is unclear if the absence of a privilege would scare them off); Note, The Lawyer-Client Privilege: Its Application to Corporations, the Role of Ethics, and its Possible Curtailment, 56 NW. U.L. REV. 235, 257 (1961) (privilege may have negligible effect).

n65. See Schwartz, supra note 60, at 683 ("There is evidence that professional and legal rules have little effect on the willingness or unwillingness of clients to talk to their lawyers."). This conclusion is supported by the way clients act toward other professionals that cannot guarantee the same degree of confidentiality. In general, laypersons visit accountants, psychiatrists, social workers, and other specially trained personnel expecting that the professional will maintain the secrecy of most information. Yet everyone is aware that, under unusual circumstances, the shield of confidentiality may be pierced. Nevertheless, clients continue to use the services of these professionals. The risk of disclosure is simply insignificant compared to the benefits the professionals can provide in dealing with the clients' needs.

n66. Most states, for example, explicitly limit the availability of discovery in criminal cases without running afoul of the constitutional guarantee that defense attorneys be "effective." See generally W. LAFAVE & J. ISRAEL, CRIMINAL PROCEDURE § 19.3(b) (detailing status of defense discovery throughout the 50 states) and authorities cited therein; see also State v. Tune, 13 N.J. 203, 210-12, 98 A.2d 881, 884-86 (1953) (rejecting liberal criminal discovery); Flannery, The Prosecutor's Case Against Liberal Discovery, 33 F.R.D. 74 (1963) (defense discovery rules place government at disadvantage). Various legal rules, including the attorney-client privilege, specifically deny attorneys access to information that may help them represent their clients.

n67. Morgan, supra note 7, at 26-27; see also Hazard, supra note 18, at 1062 (noting need to protect client's victims who seek disclosure).

n68. See Nix v. Whiteside, 475 U.S. 157, 170-71 (1986). See generally Wolfram, supra note 24.

n69. Kenneth Mann suggests that a lawyer sometimes might want a client to decide for himself an appropriate strategy and withhold information that would prevent the lawyer from presenting the claim according to that strategy, or presenting it effectively. K. MANN, supra note 59, at 17, 103-04; see also Freedman, supra note 27, at 1478-82 (lawyer may give advice that will tempt client to commit perjury). Some commentators disavow tactics by lawyers that encourage clients to manipulate the search for truth. See, e.g., Bress, Professional Ethics in Criminal Trials: A View of Defense Counsel's Responsibility, 64 MICH. L. REV. 1493, 1496-97 (1966); Noonan, supra note 27, at 1488-89. But their view of ethical conduct bears more on the "preventing misconduct" justification for confidentiality, see infra text accompanying notes 82-83, than on the issue of whether lawyers need full information to be effective.

n70. See, e.g., J. CASPER, AMERICAN CRIMINAL JUSTICE: THE DEFENDANT'S PERSPECTIVE 105-15 (1972) (criminal defendants typically perceive public defenders to be on the side of the state or, at best, a middleman); Flemming, supra note 25, at 258-60 (criminal defendants tend to mistrust and question the professional judgment of their lawyers); see also A. DERSHOWITZ, THE BEST DEFENSE 18 (1982) (regardless of defendant's assertions, "a wise defense attorney always presumes his client's guilt").

n71. Professor Burt suggests that prevalent client distrust of lawyers means that a promise of confidentiality "provides no affirmative reason to share secrets." Burt, supra note 14, at 1032. A client who believes that he and the attorney have antagonistic interests will not confide simply because of the "attorney's promise not to disclose those antagonisms to third parties." Id.

n72. See Simon, supra note 59, at 1142 (stating argument in favor of confidentiality reflects "perverse priorities" in that it shows "greater solicitude for the withholding client than for the opposing party who will be harmed by nondisclosure").

n73. Despite the rhetoric in commentaries, see, e.g., M. FREEDMAN, supra note 24, at 4-5, lawyers are not unique in their relationship to clients and their need for confidential information. See generally Moore, Limits to Attorney-Client Confidentiality: A "Philosophically Informed" and Comparative Approach to Legal and Medical Ethics, 36 CASE W. RES. 177 (1985) (analogizing legal and medical confidentiality); Noonan, supra note 27, at 1485 ("attorneys are no more essential to the conduct of general business than are accountants, bankers, and secretaries, who do not enjoy the privilege"). Of all professionals, psychiatrists and psychologists possess perhaps the greatest call for secrecy. The details they demand from patients are both intimate and critical to effective job performance. Although, like lawyers, psychiatric professionals normally may not divulge client information, there are exceptional situations in which they are permitted and even required to divulge confidences. See Tarasoff v. Regents of Univ. of Cal., 17 Cal. 3d 425, 441-42, 551 P.2d 334, 347, 131 Cal. Rptr. 14, 27 (1976) ("[T]he public policy favoring protection of the confidential character of patient-psychotherapist communication must yield to the extent to which disclosure is essential to avert danger to others.").

Even "nonprofessional" employees often need access to an employer's trade secrets or other information in order to do their jobs. Such employees, like lawyers, owe a duty of loyalty. Yet at times the law allows independent interest in disclosure to override the employer's interest in confidentiality. See, e.g., Mt. Healthy City School Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 284 (1977) (must balance first amendment rights of speaker and rights of employer in promoting efficiency of services it performs); Pickering v. Bd. of Educ., 391 U.S. 563, 574 (1968) (same).

n74. See Pepper, supra note 24, at 617 ("increasing autonomy is morally good"). But see Simon, supra note 59, at 1123-25 ("the appeal to client autonomy . . . begs the question of why the client's autonomy or right should be preferred to that of the person whose autonomy or right is frustrated by the client's activities").

n75. A few scholars have espoused the pure "hired gun" model. See, e.g., M. FREEDMAN, supra note 24, at 9-24. But most commentators believe a lawyer's loyalty to the client should not be absolute. See, e.g., Bress, supra note 69, at 1495-97; Noonan, supra note 27, at 1485; see also D. LUBAN, supra note 25, at 10; cf. Nix v. Whiteside, 475 U.S. 157, 169-71 (1986) (client may not insist upon lawyer silence about client perjury).

n76. Different lawyer-client relationships create different levels of confidence. In the context of appointed criminal representation, for example, it is often difficult to find anything remotely resembling a mutual trust relationship. Defendants tend to view their counsel as part of the "system" that is antagonistic to their interests. See, e.g., J. CASPER, supra note 70, at 106-15 (1972) (Connecticut study on attitudes of criminal defendants). In situations in which clients start from a position of mistrust, counterintuitive claims by attorneys that they will never act against the client can only serve to put the client on guard even further.

n77. See, e.g., Frankel, Search for Truth Continued, supra note 22, at 54; Wolfram, supra note 24, at 809, 867-68.

n78. Burt, supra note 14, at 1015-16 ("[Codes] obstruct[] the possibility that . . . attorneys and clients will transcend their initial mistrust. . . . [T]he codes and common attitudes of the profession may defeat aspirations for attorney-client trust by discouraging the acknowledgment of mistrust."); cf. Morgan, supra note 41, at 738 (discussing unreality of image that lawyer will never disclose); Rauh, The Lawyer's Obligation to the Public Interest, in ABA TORT AND INSURANCE PRACTICE SECTION, THE LAWYER'S PROFESSIONAL INDEPENDENCE 9, 11, 15 (1985) (discussing effect of portraying lawyers as unyieldingly allied with the client).

n79. An attorney who kowtows to improper client conduct fails to accord the client the type of respect as an equal and the type of truthful advice and representation that confidentiality is designed to foster. See S. BOK, supra note 34, at 135 (values protected by confidentiality eroded by practices of secrecy); see also Donagan, supra note 39, at 144 (questioning "dignity" argument); Rhode, supra note 7, at 608-09 (same); Subin, supra note 26, at 1160 (same).

n80. Nix v. Whiteside, 475 U.S. 157, 171 (1986).

n81. See ABA COMM'N ON EVALUATION OF PROFESSIONAL STANDARDS, MODEL RULES OF PROFESSIONAL CONDUCT Rule 1.6 comment (May 30, 1981 draft). One economist, however, questions whether encouraging legal advice and triggering the adversary system will in fact lead to socially desirable results. Shavell, supra note 58. Professor Shavell, however, defines undesirable acts according to whether the party's gain from an act would be less than the harm done. Id. at 131. Under this framework, what society deems to be misconduct -- even criminal misconduct -- may not fall to the "undesirable" level.

n82. Under Professor Shavell's economic model, protection of confidentiality will in many circumstances "have no effect on a party's decision to obtain . . . advice." Shavell, supra note 58, at 130. But see id. at 14-15, 18 (noting circumstances in which confidentiality rules may have an effect).

n83. As Professor Rhode notes, "[L]ittle is known about the extent to which lawyers have managed to channel patrons along 'proper paths.'" Rhode, supra note 7, at 615; see also Subin, supra note 26, at 1166-73 (questioning dissuasion of clients from wrongdoing as a justification for absolute confidentiality); Weissenberger, Toward Precision in the Application of the Attorney-Client Privilege for Corporations, 65 IOWA L. REV. 899, 899 (1980) (it is unclear that increased client candor results "in a greater probability of lawful client behavior").

Lawyers participating in the Tompkins County study clearly believed confidentiality works to deter client wrongdoing. Of the lawyers responding, 43.5% claimed that "the existence of attorney-client confidentiality [occasionally or frequently] enabled [them] to obtain information and then dissuade a client from taking improper action." An additional 33.9% stated that confidentiality had enabled them to dissuade improper action, but only "rarely." T.C. Table L1, at 11.

n84. See generally M. FREEDMAN, supra note 24; S. LANDSMAN, supra note 32; Alschuler, The Preservation of a Client's Confidences, supra note 22; Frankel, Umpireal View, supra note 22, at 1031; Fuller, The Adversary System, in H. BERMAN, TALKS ON AMERICAN LAW 34 (1971). Rarely has anyone considered the effect of professional norms on attorneys themselves. See generally Flynn, Professional Ethics and the Lawyer's Duty to Self, 1976 WASH. U.L.Q. 429, 435 (CPR does not adequately define the lawyer's professional duty to self); Wasserstrom, supra note 27, at 8 (introducing notion that professional codes of ethical responsibility produce "amoral" conduct by lawyers).

n85. See Levine, supra note 62, at 811-12 (discussing codified exceptions to attorney client confidentiality when disclosure relates to accusation against attorney or when attorney's compensation is at issue).

n86. See MODEL RULES Rule 1.6(b); CPR DR 4-101(C)(4).

n87. CPR DR 4-101(C)(4); accord MODEL RULES Rule 1.6(b)(2) ("A lawyer may reveal . . . information to the extent the lawyer reasonably believes necessary . . . to establish a claim . . . in a controversy [with the client].").

n88. Id.; see also Meyerhofer v. Empire Fire & Marine Ins. Co., 497 F.2d 1190, 1194-95 (2d Cir. 1974), cert. denied, 419 U.S. 998 (1974).

n89. See supra note 5 and accompanying text.

n90. Traditionally, privilege exceptions have been incorporated into confidentiality rules. See, e.g., CPR DR 4-101(C)(2); MODEL RULES Rule 1.6 comment ("The lawyer must comply with the final orders of a court . . . requiring the lawyer to give information about the client.").

n91. Courts have applied the attorney-client privilege from the perspective that it "has such an effect on full disclosure of the truth that it must be narrowly construed." Underwater Storage, Inc. v. United States Rubber Co., 314 F. Supp. 546, 547 (D.D.C. 1970); see also Weil v. Investment/Indicators, Research & Management, Inc., 647 F.2d 18, 24 (9th Cir. 1981); Diversified Indus. v. Meredith, 572 F.2d 596, 602 (8th Cir. 1977); Radiant Burners, Inc. v. American Gas Ass'n, 320 F.2d 314, 323 (7th Cir.) (en banc), cert. denied, 375 U.S. 929 (1963). They have stretched client waiver notions to cover a myriad of situations in which clients do not intend to disclose confidential communications. 2 J. WEINSTEIN & M. BERGER, EVIDENCE P511[02], at 511-6-7 (1986); see, e.g., In re Victor, 422 F. Supp 475, 476 (S.D.N.Y. 1976) (client waives privilege by placing documents in a box outside attorney's office door); see also Blackburn v. Crawfords, 70 U.S. (3 Wall.) 175, 194 (1865); In re Grand Jury Investigation of Ocean Transp., 604 F.2d 672 (D.C. Cir. 1972), cert. denied, 444 U.S. 915 (1979); Agnew v. State, 51 Md. App. 614, 649-51, 446 A.2d 425, 444-45 (Md. Ct. Spec. App. 1982) (statement in Vice President Agnew's autobiography that he had "told his lawyer of his innocence" waived privilege); cf. Von Bulow v. Von Bulow, 114 F.R.D. 71 (S.D.N.Y. 1987) (waiver through lawyer's publication of book), aff'd in part and rev'd in part, 828 F.2d 94, 101 (2d Cir. 1987) (client implicitly authorized and encouraged the publication).

Judges have expanded the doctrine that a client gives up the privilege when she shares the information. Although some modern courts have focused on whether the client negligently participated in a waiver, see, e.g., Transamerica Computer Co. v. Int'l Business Mach. Corp., 573 F.2d 646, 652 (9th Cir. 1978), Hartford Fire Ins. Co. v. Garvey, 109 F.R.D. 323, 329 (N.D. Cal. 1985), Mendenhall v. Barber-Greene Co., 531 F. Supp. 951 (N.D. Ill. 1982), others have not. Courts have ordered disclosure when third parties have stolen documents, see 8 J. WIGMORE, supra note 18, § 2326, at 633-34, eavesdropped surreptitiously, see, e.g., Clark v. State, 159 Tex. Crim. 187, 191-93, 261 S.W.2d 339, 342, cert. denied, 346 U.S. 855 (1953), Dobbins v. State, 483 P.2d 255, 260 (Wyo. 1971), Schwartz v. Wenger, 267 Minn. 40, 43, 124 N.W.2d 489, 492 (1963), or uncovered the information by rifling through the client's abandoned trash. Suburban Sew 'N Sweep, Inc. v. Swiss-Bernina, Inc., 91 F.R.D. 254, 258 (N.D. Ill. 1981). But see Davidson & Roth, Waiver of the Attorney-Client Privilege, 64 OR. L. REV. 637, 640-41 (1986) (privileged communications should be inadmissible in evidence when they become known to third party through theft, misappropriation, or the inadvertence of the lawyer); see also ALI, MODEL CODE OF EVIDENCE Rule 210 comment b (1942); cf. UNIFORM RULES OF EVIDENCE, Rule 26 comment (1953) (noting modern trend contrary to Wigmore's position regarding theft of documents). Judges also have been liberal in holding that attorney errors forfeit clients' right to confidentiality, even when the clients have not been consulted in the decision to give up the right. See, e.g., United States v. El Paso Co., 682 F.2d 530, 541 (5th Cir. 1982), cert. denied, 466 U.S. 944 (1984); Perrignon v. Bergen Brunswig Corp., 77 F.R.D. 455, 460-61 (N.D. Cal. 1978); Bendele v. Tri-County Farmer's Co-Op, 635 S.W.2d 459, 464 (Tex. Ct. App.), aff'd in part, vacated in part on other grounds, 641 S.W.2d 208 (Tex. 1982).

n92. ABA, CANONS OF PROFESSIONAL ETHICS Canon 15 (1908).

n93. The CPR notes limited circumstances in which lawyers have an option to withdraw from a client's representation, DR 2-110(C), and describes narrow exceptions to rules forbidding attorney publicity and solicitation, DR 2-101 to -103, trial publicity, DR 7-107, and disclosure of client confidences. DR 4-101. The Model Rules provide somewhat more leeway for the exercise of attorney discretion, but again only in limited circumstances. See, e.g., Rules 1.2(c) (lawyer may limit objectives of representation), 1.6(b) (exceptions to confidentiality), 1.13 (corrective measures within organizational client), 1.14 (disabled client), 1.16(b) (withdrawal), 2.1 (right to give "moral" advice), 6.3-6.4 (participation in legal services and law reform activities), 7.2-7.5 (advertising); see also Rhode, supra note 7, at 601 (discussing role of discretion in the Model Rules). For an excellent compilation of authorities and discussion of the view that attorney independence has declined over time, see Gordon, supra note 24, at 48-51. Professor Gordon notes that the concept of "independence" from client control means different things to different people, id. at 6-10, and is considered an evil by some. Id. at 68-83.

n94. Flynn, supra note 84, at 436.

n95. In describing the corporate bar's willingness to "acquiesce" uncritically in client plans, Professor Gordon notes a paradox produced by the tradition of amorality: "[Lawyers] belong to a profession in which higher status correlates with diminished autonomy. Influence varies inversely, subservience directly, with professional standing. . . . Even when top management would prefer that their lawyers serve as the corporate superego, there is a deliberate shrinking from that role." Gordon, supra note 24, at 57-58.

n96. Flynn, supra note 84, at 444; see also Landesman, supra note 42, at 785-86 (in deciding whether to disclose information that an attorney is legally entitled to disclose, he also must consider any personal obligations).

n97. Wasserstrom, supra note 27, at 5. For good discussions of the effects professional amorality may have on the lawyer's individual character traits, see Eshete, supra note 27, at 274-76, 279; Wolf, Ethics, Legal Ethics, and the Ethics of Law, in D. LUBAN, supra note 25, at 38, 52.

n98. The lawyer may refuse to participate in misconduct or fraud upon a tribunal. She can suggest moral conduct to the client. See CPR DR 7 102(B)(1). The lawyer can attempt to manipulate the facts to bring them within one of confidentiality's limited disclosure exceptions. See supra note 3. In some situations, the lawyer may threaten to withdraw. See CPR DR 2-110; MODEL RULES Rule 1.16.

n99. Proponents of role differentiation argue that lawyers must adopt a professional self-image that makes moral conduct that would be immoral outside the role. See generally Donagan, supra note 39, at 123-49; Pepper, supra note 24; Wolf, supra note 97, at 38-59. Other commentators have argued that moral principles cut across artificial roles. See, e.g., A. GOLDMAN, supra note 25, at 133-48.

Of course, even the underlying notion that in our legal system every client is entitled to a zealous advocate may be question-begging, particularly outside of the criminal context. If our system does guarantee clients lawyers in civil cases, but see Lassiter v. Department of Social Servs., 452 U.S. 18, 26-27 (1981) (strong presumption against appointing counsel in civil cases), it does not guarantee anyone access to a particular lawyer's services. See Rauh, supra note 78, at 15-16 (lawyer's duty to client may yield to higher interest in the administration of justice).

n100. See Simon, The Ideology of Advocacy: Procedural Justice and Professional Ethics, 1978 WIS. L. REV. 30, 124-30 ("In a setting where a party was conscious of his responsibility to press his claims in terms of substantive justice, . . . [i]t would always be possible that he might be led . . . to take a different view of his original claim. Procedural justice . . . [forces] litigants into roles of stylized aggression.").

Interestingly, one of the asserted goals of absolute confidentiality rules is to enable lawyers to influence clients to "do the right thing." Yet to the extent exceptions force lawyers to think about moral issues or enhance clients' respect for their attorneys, the exceptions rather than the basic rule may produce the lawyer's ability to dissuade. Cf. Burt, supra note 14, at 1032-34 (imposing negligence standard for disclusure would force lawyers to engage in moral dialogues with their clients and would enhance attorney-client trust).

n101. See C. WOLFRAM, supra note 18, § 6.1, at 247 (lawyers make the laws concerning privilege and benefit most from them); Wolfram, supra note 24, at 809, 838-39 (rules allow attorney detachment and innocence).

n102. Professor Leubsdorf notes that an emerging branch of professional responsibility scholarship calls on lawyers to "stop hiding behind rules, roles, and institutions and to take responsibility for their actions." Leubsdorf, supra note 16, at 1045-52. This Article does not advocate the pure "personal responsibility model" Professor Leubsdorf describes, which disavows any role differentiation and generally opposes professional regulation. Id. at 1047. I simply suggest that rules that afford greater opportunity for lawyers to exercise moral judgments may serve society well.

n103. See Leubsdorf, supra note 16, at 1046 ("the traditional system encourages the lawyer to pursue his client's supposed interests to the limit . . . without taking responsibility for the results"). The recent "professional responsibility" literature has begun to focus on the negative consequences of overzealous litigation. See, e.g., ABA, ATTACKING LITIGATION COSTS AND DELAY 1, 5 (1984) ("citizens' access to justice . . . is threatened"); Janofsky, Reducing Court Costs and Delay: An Overview, 16 U. MICH. J.L. REF. 467 (1983) (discussing existence and ramifications of costly, lengthy litigation). Commentators have exhibited particular concern over abusive discovery techniques that obscure the facts and tax opponents' resources, see, e.g., ABA COMM'N ON PROFESSIONALISM, IN THE SPIRIT OF PUBLIC SERVICE: A BLUEPRINT FOR THE REKINDLING OF LAWYER PROFESSIONALISM 5 (1986) (discussing "'scorched earth' strategy of litigation"); Brazil, Civil Discovery: How Bad Are the Problems, 67 A.B.A.J. 450-52 (1981); Flegel, Discovery Abuse: Causes, Effects, and Reform, 3 REV. LITIG. 1, 21-22 (1982), and nonmeritorious lawsuits attorneys file and maintain in order to obtain settlements, see, e.g., L. BEBCHUK, SUING SOLELY TO EXTRACT A SETTLEMENT OFFER (1986) (discussion paper) (economic analysis); S. PEPE, STANDARDS OF LEGAL NEGOTIATIONS: INTERIM REPORT FOR ABA COMMISSION ON EVALUATION OF PROFESSIONAL STANDARDS AND ABA HOUSE OF DELEGATES 254-55 (1983) (empirical study discussing willingness of attorneys to negotiate settlement on basis of client falsehoods); Rhode, supra note 7, at 599 ("the prime objective in much contemporary litigation is to force the adversary to 'cry craven' well before . . . adjudication").

n104. See, e.g., G. HAZARD, JR., ETHICS IN THE PRACTICE OF LAW 146 (1978); Donagan, supra note 39, at 132; Landesman, supra note 42, at 776, 785-86; Noonan, supra note 27, at 1491-92.

n105. See, e.g., G. HAZARD, JR., & W. HODES, THE LAW OF LAWYERING XXXV (1986); Fried, The Lawyer as Friend: The Moral Foundations of the Lawyer-Client Relation, 85 YALE L.J. 1060, 1086 (1976).

n106. This view is consistent with the general orientation of the professional codes. The CPR and the Model Rules, with very limited exceptions, talk in terms of attorneys' obligations -- to clients, the courts, and society as a whole. The CPR is divided into nine broad sections of what lawyers "should" do. The Model Rules start with a preamble describing "a lawyer's responsibilities" and proceeds with a series of "shall's" and "shall not's." Neither modern code elaborates on attorneys' own interests in practicing law nor details their rights vis-a-vis clients.

The reasons for this may have to do with the spotlight focused upon the legal profession when it enacts or reforms ethical norms. See Gillers, supra note 42, at 244 (journalists are watchdogging the profession). Proceedings surrounding the promulgation of the 1969 CPR and 1983 Model Rules, for example, drew substantial comment in both the trade and popular press. See, e.g., D. LUBAN, supra note 27, at 180-81; Taylor, Lawyers Vote Against Disclosure of Fraudulent Activity by Clients, N.Y. Times, Feb. 8, 1983, at A1, col. 2; Taylor, Ethics and the law: A Case History, N.Y. Times, Jan. 9, 1983, § 6 (Magazine), at 31; Whose Lawyer?, Wall St. J., Aug. 15, 1983, at 16, col. 1; see also G. HAZARD, JR., & W. HODES, supra note 105, at 88 (discussing media attention to proposed confidentiality rules). Openly self-serving self regulation would invite public criticism that might, in the long run, prompt government intervention. The preamble to the Model Rules makes explicit the bar's desire to avoid outside regulation. See also ABA SPECIAL COMMITTEE ON EVALUATION OF DISCIPLINARY ENFORCEMENT, PROBLEMS AND RECOMMENDATIONS IN DISCIPLINARY ENFORCEMENT 2, 8-9 (1970) (unless the profession "reform[s] its own disciplinary structure," those "outside the profession" will impose changes); cf. Gillers, supra note 42, at 273-74 (questioning bar's justifications for avoiding regulation); Wolfram, Barriers to Effective Public Participation in Regulation of the Legal Profession, 62 MINN. L. REV. 619, 623 (1978) (outside pressures for regulation may "threaten the bar's traditional self-regulation unless the legal profession takes effective action").

n107. Rauh, supra note 78, at 15.

n108. See Subin, supra note 26, at 1164 (noting confessional aspect of lawyer-client communications).

n109. See CPR DR 7-102(A)(1); MODEL RULES Rule 3.1.

n110. Of course, nothing other than the tradition engendered by the longstanding professional codes prevents lawyers from telling clients up front that there are limits to what they will hear or do. See Rauh, supra note 78, at 17-18; see also MODEL RULES Rule 1.2(c).

n111. Professor Post recounts a joke that reflects this widely-shared view:

Question: Why did the research scientist substitute lawyers for rats in his laboratory experiments?

Answer: Lawyers breed more rapidly, scientists become less attached to them, and there are some things that rats just won't do.

Post, On the Popular Image of the Lawyer: Reflections in a Dark Glass, 75 CALIF. L. REV. 379, 379 (1987).

As Professor Post suggests, the reasons for this view of lawyers and its validity are complex. Id. at 386-89. The profession must constantly balance its responsibility to clients on the one hand and society or the law on the other. Yet the negative public perception of lawyers may still tell us that the profession currently strikes a faulty balance.

n112. Thurman, supra note 7, at 19; cf. Gordon, supra note 24, at 20-21:

If clients, including those who prefer to be law-abiding even when nobody is likely to know when they are not, habitually consult lawyers who recommend only the most literal forms of compliance [with the law] and widen every loophole far enough to lawyers to disclose when "the lawyer 'reasonably deems it to be necessary in the public interest'" would affect "public perception of the integrity of lawyers." Of the 61 respondents who thought the change would have any effect, nearly two thirds (65.6%) believed loosening confidentiality would make the public perceive lawyers in a "more favorable light." Among all respondents, 21 (23.3%) thought public perception would become less favorable, 40 (44.4%) expected an improvement, and 29 (32.2%) anticipated no change. T.C. Table J1A, at 91.

n113. See Note, Client Fraud and the Lawyer -- An Ethical Analysis, 62 MINN. L. REV. 89, 117 (1977) (for legal profession to vindicate claim of integrity, it must "convince the public that there are certain things a lawyer cannot and will not do, even for his client").

n114. Id. at 117-18. One author speculates that ambiguity in well-intentioned professional standards may be more damaging to public confidence than clear self-serving rules. Note, Attorney Client Confidentiality: A New Approach, 4 HOFSTRA L. REV. 685, 688 (1976).

n115. See, e.g., authorities cited supra note 7.

n116. See, e.g., G. HAZARD, JR. & W. HODES, supra note 105, at 89 ("Although there is no empirical evidence of the precise degree to which clients rely on the principle of confidentiality, it is intuitively obvious that lawyers will be better able to gain the trust of clients, to serve them, and to help them obey the law under a requirement of confidentiality that goes beyond mere tradition.").

n117. Cf. Empirical Research Project, Corporate Legal Ethics -- An Empirical Study: The Model Rules, the Code of Professional Responsibility, and Counsel's Continuing Struggle Between Theory and Practice, 8 J. CORP. L. 601, 621-29 (1983) (empirical study of, inter alia, the operation of attorney-client privilege with respect to employee information).

n118. The functions of attorneys and other professionals often overlap. For example, lawyers must at various times act as accountants, business advisors, social workers, psychologists and psychiatrists, marriage counselors, and realtors. Similarly, other professionals -- especially accountants and realtors -- are regularly called upon to give legal advice relating to transactions within their fields of expertise.

n119. Note, supra note 26, at 1227. There are several reasons to caution against too much reliance on the Yale study. First, the number of subjects was limited. Second, subjects were not chosen with sufficient randomness to assure the significance of the results. See id. at 1227 n.6. Third, the study was conducted by a legal periodical, without statistical rigor. Still, the study contains the only available empirical data. The results provide valuable food for thought.

n120. Id. at 1236.

n121. Id. at 1232. Ninety of 102 responding attorneys believed that the privilege helped induce disclosure, while only 55 of 92 laypersons agreed. Id. at 1232 n.38.

n122. Id. at 1239 n.81.

n123. Of 108 laypersons responding, 32 thought that attorneys would disclose and another 21 did not know. Id. at 1262, question (5).

n124. Id. at 1262, question (7). Thus, only 49 of 108 affirmatively opposed enforced disclosure.

n125. Id. at 1236 n.59 (55 of 108 believed elimination would deter disclosures). The study did not consider the effect of simply creating additional exceptions to the antidisclosure principles.

n126. Id. at 1232.

n127. Id. at 1262, question (6).

n128. Id. at 1255.

n129. Id. at 1262, question (5).

n130. Id. at 1262, question (6).

n131. Id. at 1262, question (7).

n132. Lawyers were perceived as slightly less likely than other professionals to "repeat to others matters . . . told them in confidence." Id. at 1262, question (4). But when asked whether eight different professions could legally resist disclosing confidential information in court, lay subjects ranked lawyers squarely in the middle. Id. at 1262, question (5).

n133. Virtually all American jurisdictions recognize the attorney-client privilege. At the time of the Yale study, several states provided a psychiatrist-patient privilege, see id. at 1252, fewer protected psychologists, social workers, and marriage counselors, see id. at 1252-53, fewer yet limited accountant disclosures, see id. at 1247, and none had privileges applying to business advisors and realtors. See id. at 1251. Societies representing many professions do maintain ethical codes forbidding disclosure. See R. GORLIN, CODES OF PROFESSIONAL RESPONSIBILITY 133 (1986) (psychiatry); id. at 143 (psychology); id. at 161 (social work); id. at 239 (accounting). But all of these codes allow disclosures in the context of judicial proceedings and some include exceptions which in practice open up the possibility of more frequent revelations. See, e.g., id. at 165 (social worker may disclose "for compelling professional reasons"); id. at 143 (psychologist may disclose when "not to do so would result in clear danger to the person or others").

n134. See Moore, supra note 73, at 196-211 (comparing confidentiality in medical and legal professions).

n135. Rhode, supra note 7, at 614 ("From a historical and cross-cultural perspective, it appears that most professionals . . . have managed to discharge confidential counseling without the absolute freedom from third-party obligations that the organized bar now claims") and authorities cited at n.93.

n136. In an attempt to be comprehensive, I contacted all 145 attorneys listed with the Tompkins County Bar Association, in Martindale-Hubbell, and in the local Yellow Pages. Many of the attorneys so listed were not practicing attorneys, had moved from the jurisdiction, or had died.

n137. Of the lawyers who responded, the two largest groups, 50% and 22.6% respectively, engaged primarily in general practice and real estate. The remaining 27.4% engaged in a variety of practices including corporate (6.5%), poverty or public interest law (6.5%), tax, trusts, and estates (4.8%), and family/matrimonial law (4.8%). Consistent with this Article's limited scope, see supra text accompanying notes 24-27, only one of the respondents (1.6%) engaged primarily in criminal practice. T.C. Table L1, at 2.

n138. More than two-thirds of the subjects were over 30 years old. T.C. Table J1A, at 3 (68.6%). Most had completed their high school educations. T.C. Table J1A, at 6 (98.1%). Nearly two-thirds had a college or higher degree. T.C. Table J1A, at 6 (63.8%).

n139. My anecdotal observation over the years has been that the mock jury panels represent a fairly good cross-section of the Tompkins County community, except for their racial mix. According to the survey responses, 96.2% of the responding laypersons were white. T.C. Table J1A, at 5. Two-thirds were female. T.C. Table J1A, at 4.

n140. T.C. Table L2, at 40 (85%). Throughout my discussion of the study results, the percentages refer only to those lawyers, clients, and laypersons who responded to the particular survey question. Those who answered other questions, but responded to the particular query with an "I don't know," "I don't remember," or by ignoring the query entirely were excluded from the result calculations for that query.

n141. T.C. Table J1A, at 21 (52.5%). It is, however, impossible to know how the lay subjects would react in a real-life setting. These respondents were answering in a vacuum of information. Half of the subjects who threatened to withhold information were operating on the incorrect assumption that current rules require absolute confidentiality. T.C. Table J7A, at 102.

n142. T.C. Table J1A, at 15 (29.6%).

n143. RESTATEMENT (SECOND) OF AGENCY § 395 comment f (1957); see also supra text accompanying notes 48-49.

n144. T.C. Table J5A, at 13.

n145. T.C. Table L1, at 11 (77.4%).

n146. Over 56% never or "rarely" used confidentiality in this way. Id. Of the rest, 40.3% have dissuaded misconduct "occasionally"; only 3.2% have done so "frequently." Id.

n147. T.C. Table L1, at 12 (37.7%).

n148. T.C. Table L1, at 12.

n149. T.C. Table L16, at 59.

n150. As might be expected, the study showed that lawyers practicing the areas of law in which problems of confidentiality arise infrequently (e.g., real estate) inform clients more rarely than other attorneys. The study, however, also reveals that the lack of explanation appears across the board. See T.C. Table L18, at 4-5.

Lawyers practicing specialties in which confidentiality plays less of a role -- for example, real estate and patent law -- may understandably see no need to explain confidentiality rules as a routine. But for these lawyers, strict rules also cannot be justified on the same instrumental basis as for others. Their failure to inform clients of the rules may itself suggest that code drafters should distinguish among practitioners in crafting exceptions.

It is also important to note that lawyers practicing the areas of law rarely involving confidentiality issues may carry their understatement of confidentiality too far. Real estate clients, for example, may well wish to keep secret many kinds of personal financial information. Nevertheless, one real estate attorney, in responding that he would not complete the survey, stated that "since I no longer practice litigation, your survey does not seem to apply to my work." T.C. ID No. L64.

n151. T.C. Table L1, at 4.

n152. Id.

n153. T.C. Table L1, at 5.

n154. Id.

n155. T.C. Table J1A, at 10. To avoid confusion due to clients who had consulted multiple attorneys who informed them of confidentiality differently, the survey asked the subjects to identify what their first attorneys told them. In a separate question, the survey asked clients the more general question, "How many of the attorneys you consulted told you that information about your case would be kept confidential?" A slightly lower percentage, 53.5%, responded that none of their attorneys had mentioned confidentiality. T.C. Table J1A, at 8.

n156. T.C. Table J1A, at 13.

n157. T.C. Table J1A, at 12.

n158. Id.

n159. T.C. Table J5A, at 4.

n160. Id.

n161. Id.

n162. T.C. Table J3A, at 6 (42.4%).

n163. T.C. Table J3A, at 7. That figure is, however, raised dramatically to 64.8% by excluding cases in which a court orders an attorney to disclose. Id.

n164. T.C. Table J3A, at 7. An additional 5.7% envisioned disclosure in other cases. Id. Reasonable minds might disagree on the significance of these findings. While they do show that many laypersons believe lawyers will follow their own moral instincts rather than the ethical rules, most laypersons seem to believe lawyers will obey the professional standards.

n165. 63.6% and 68.7% respectively. T.C. Table J3A, at 14, 15.

n166. 42.4% and 39.7% respectively. T.C. Table J3A, at 12, 16.

n167. See generally T.C. Table J3A, at 18-23.

n168. This table is a composite of T.C. Tables L1, at 8-9; J3A, at 12-17; and J3A, at 18-23.

n169. Lawyer and client responses to questions about dealings with priests provide an interesting insight into the limitations of confidentiality rules in shaping clients' actions vis-a-vis professionals. In most states, the law does not require priests to keep information confidential. Nor does it protect them against revealing information in litigation. Nevertheless, priests routinely keep confessions confidential, at pain of imprisonment. Despite the absence of a priest rule, most of the lawyers surveyed, 81%, correctly concluded that clients don't believe lawyers protect confidences better than priests. T.C. Table L1, at 8. Of the clients, only 22.2% believed attorneys are legally more required to preserve confidentiality. T.C. Table J3A, at 13. Even fewer, 14.3%, believed lawyers in practice preserve confidentiality better. T.C. Table J3A, at 19.

n170. This table is a composite of T.C. Tables J3A, at 12-17 and J3A, at 24-29.

n171. T.C. Table L1, at 14 (85%).

n172. T.C. Table L1, at 5 (71.9%). This figure might be explained on the basis that lawyers assume clients know about confidentiality without being told. To some extent, the survey supports that conclusion. A large majority of clients whose lawyers never explained confidentiality, 91.1%, believed confidentiality exists, though 41.2% believed it exists to a greater extent than the law actually requires. T.C. Table J4A, at 86-87. On the other hand, a significant number of clients believed similar confidentiality guarantees exist for dealings with priests, doctors, psychologists and psychiatrists, accountants, and social workers. See T.C. Table J3A, at 12-17. It is probably fair to conclude that, in the absence of an explanation of what attorney-client confidentiality means, clients are at least somewhat confused.

n173. T.C. Table L1, at 5. Similarly, lawyers do not seem to share the code drafters' concern with client lies or withholding of information. Few of the lawyers surveyed believed their clients lied in more than a quarter of their cases. T.C. Table L1, at 10 (23.4%). When, however, lawyers thought their clients lied or withheld information, only 19.3% of the lawyers believed the failure to disclose significantly affected more than 25% of the cases. T.C. Table L1, at 10.

n174. Of all clients surveyed, 11.3% withheld information from their lawyers. T.C. Table J1A, at 16. Four of 34 clients (11.8%) who were not told of confidentiality withheld information. T.C. Table J4A, at 83.

n175. T.C. Table J1a, at 21 (52.5%).

n176. I.e., 28.8% of the total 52.5% who initially said they would withhold. T.C. Table J1a, at 21, 23.

n177. T.C. Table J1A, at 16.

n178. T.C. Table L1, at 5.

n179. T.C. Table L1, at 5.

n180. T.C. Table J1A, at 10 (1.7%).

n181. Interestingly, an even higher percentage of lawyers who usually do not explain confidentiality, or do so only when they perceive a problem, tend to overstate confidentiality. Of the lawyers who tell clients of confidentiality in less than half their cases, 84.6% say only that "all communications are confidential." T.C. Table L7, at 27. Lawyers responding to client hesitation or a problem of confidentiality -- those who would be expected to take particular care to note the exepttions for fear of misleading the client -- overstate confidentiality 82.8% of the time. T.C. Table L10, at 63.

n182. T.C. Table L1, at 7.

n183. T.C. Table L12, at 26.

n184. T.C. Table L1, at 8.

n185. T.C. Table L12, at 28.

n186. T.C. Table J12A, at 7. Approximately 48.6% of the clients correctly understood the general concept that lawyers must usually keep client communications confidential, but that there are specific exceptions allowing disclosure. T.C. Table J12A, at 7. Interestingly, the clients who misperceived that confidentiality is absolute did not necessarily embark on the attorney-client relationship with that understanding. Only 55.6% of those clients had that idea before consulting an attorney. T.C. Table J13A, at 9.

n187. Because Tompkins County lawyers tend to exaggerate the scope of confidentiality, clients in fact seemed to rely on confidentiality more than they should have. Of the surveyed clients, 29.6% stated they gave information to their attorney(s) that "they would not have given without a guarantee of confidentiality." T.C. Table J1A, at 15. Half of these clients, however, were operating under the incorrect belief that confidentiality is absolute. T.C. Table J5A, at 4. Cf. generally Dan-Cohen, Decision Rules and Conduct Rules: On Acoustic Separation in Criminal Law, 97 HARV. L. REV. 625 (1984) (arguing that society may be better off under system in which the general public does not understand the substantive criminal law).

n188. Consider a situation in which a lawyer feels the need to invoke an exception to confidentiality; for example, when a client tells the attorney he intends to commit perjury. If the lawyer previously told the client that all communications are confidential, it is difficult to imagine how the lawyer-client relationship can continue once the lawyer threatens to disclose the perjury. In precisely such a situation, the court of appeals in Whiteside v. Scurr, 744 F.2d 1323, reh'g denied, 750 F.2d 713 (8th Cir. 1984), found that the lawyer's threat destroyed the attorney-client relationship and therefore rendered the subsequent representation ineffective. Id. at 1328. The Supreme Court, in its haste to rule that the defendant had no right to perjure himself, essentially ignored the importance of the attorney-client trust relationship. See Nix v. Whiteside, 475 U.S. 157, 169-71 (1986) (attorney does not have to tolerate passively clients' perjury proposals).

n189. T.C. Table L1, at 16.

n190. T.C. Table L1, at 16.

n191. A significant number of laypersons stated that a public interest exception would cause them to consult attorneys less frequently, T.C. Table J1A, at 88 (39.1%), and trust them less. T.C. Table J1A, at 89 (30.8%); cf. Table J1A, at 88 (if exception adopted, 2.2% would never consult an attorney). Some of their responses may be attributable to the fact that many of the respondents harbored the belief that present confidentiality requirements are absolute. T.C. Tables J6A, at 42 and J10A, at 4 (52.5% of laypersons responding that they would consult attorneys less); J6A, at 47 and J10A, at 9 (38.5% of laypersons responding that they would trust attorneys less).

n192. T.C. Table J1A, at 91.

n193. Id.

n194. Id.

n195. A caveat is in order. The questionnaire in this instance was worded poorly. The lay respondents' opinions of how others might perceive a change in the rules may be no more reliable than the lawyers' or the reader's. Asking how the respondents themselves would perceive the change would have gauged lay attitudes more accurately. I offer the reflections in the text nonetheless, because respondents' views of others' reactions probably provides some insight into the respondents' own views.

n196. See supra text accompanying note 145.

n197. See supra text accompanying note 146.

n198. T.C. Table L1, at 13.

n199. Id.

n200. For example, because the lawyer would discover the information anyway or the client would ultimately decide to behave on his own.

n201. The Tompkins County questionnaires asked about hypotheticals (13)-(15). The results are not included in the above discussion or the tables because hypothetical (13) in substance duplicates the defective airplane example and because the status of hypotheticals (14) and (15) under New York ethical rules is questionable. In general, both lawyer and layperson responses were in line with the answers relating to the other questions.

n202. This Article also refrains from endorsing any specific exceptions, although it does suggest that more exceptions may be needed.

n203. The questionnaire did not ask clients where the representation occurred. Nevertheless, it is reasonable to assume, for purposes of this analysis, that Tompkins County residents are likely to use local counsel for their legal problems.

n204. See, e.g., N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 1-102(A)(3)-(5), 7-101(B)(2), 7-101(A)(4), (6), (7), 7-103(B)(1)-(2) (1978).

n205. See id., DR 7-102(B)(1) (authorizing disclosure of fraud "except when the information is protected as a confidence"). In contrast, Alaska, New Jersey, Maryland, Virginia, and Wisconsin, among others, permit disclosure of client confidences to prevent certain types of fraud. Other jurisdictions make an exception for frauds upon a tribunal. See supra note 5.

n206. N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 4-101 (1978).

n207. Arguably, in the case of the mentally imbalanced and potentially dangerous client, the lawyer may be attempting to prevent a future crime. However, since the lawyer does not know his client has formed any "intention" to commit a crime, DR 4 101(C)(4) does not seem applicable.

n208. See N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 2-110(C)(1)(e) (1978); see also N.Y.S. Bar Association Committee on Professional Ethics, Op. 562 (July 19, 1984) (setting guidelines for actions attorney should take when a client threatens to commit a crime).

n209. For example, if the lawyer can determine that her own client is involved in the kidnapping in hypothetical (1) or that the clients in hypotheticals (4) and (6) are continuing to engage in continuing criminal fraud, the lawyer might be able to disclose on the basis of the crimes exception. In the narrow and carefully defined hypotheticals presented to the study subjects, however, no factual escape hatches were included.

n210. N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY DR 7-102(B)(2) (1978).

n211. Id., DR 7-102(B)(1).

n212. For example, if the house sale closes and the buyer subsequently learns of the fraud, the client may become liable for contract or tort damages. In the perjury and benefits examples, the client is exposed to criminal prosecution.

n213. Such an extreme interpretation of the waiver doctrine would unduly muzzle clients who are accused of wrongdoing. For example, it would prevent accused defendants from asserting their innocence in any setting, including employment interviews or conversations with their families. Nevertheless, a few courts have taken such an extreme view of waiver in applying the privilege. See, e.g., Agnew v. State, 51 Md. App. 614, 446 A.2d 425, 44-54 (1982) (publication of Agnew's book, with detailed conversations between him and former attorney, constituted waiver); Von Bulow v. Von Bulow, 114 F.R.D. 71, 76-77 (S.D.N.Y. 1987) (attorney's publication of book disclosing communications with client constituted waiver because client had encouraged and acquiesced in publication).

n214. See generally N.Y.S. BAR ASSOCIATION, THE LAWYER'S CODE OF PROFESSIONAL RESPONSIBILITY EC 4 (1978); see also id., EC 4-4 ("The attorney-client privilege is more limited than the ethical obligation of a lawyer to guard the confidences and secrets of the client.").

n215. Earlier, this Article confined its own scope to the noncriminal context. See supra text accompanying notes 24-27. In most of the hypotheticals, the lawyer represents the client only with respect to civil matters. Only in the scenario involving the Vice President who denies wrongdoing is the lawyer identified as criminal defense counsel.

n216. See, e.g., N.J. RULES OF PROFESSIONAL CONDUCT Rule 1:6(b)(1) (lawyer must disclose to prevent client "from committing a criminal, illegal or fraudulent act that the lawyer reasonably believes is likely to result in death or substantial bodily harm or substantial injury to the financial interest or property of another"); WIS. SUPREME COURT RULES Rule 20:1.6(b) (same).

n217. To be fair, it is important to note that most of the responding attorneys answered the questions without research. If confronted by the hypothetical situations in real life, one would hope and expect that the attorneys would investigate whether the code allows disclosure or even seek an opinion from the bar. Still, the responding attorneys did believe they had a basis for their answers. They did not choose the offered response, "I really couldn't hazard a guess."

n218. The answers of laypersons as a whole regarding the hypotheticals closely paralleled those of clients. T.C. Table J1A, at 42-83. The following discussion thus refers only to the client responses.

n219. See Table IV, infra text accompanying note 225.

n220. See Table V, infra text accompanying note 226.

n221. The survey phrased the question in terms of the "good attorney" rather than the specific responding attorney in order to obtain the most honest possible answers. Although the author promised responding attorneys anonymity and took steps to assure that he could not identify individual respondents, individual lawyers still might have feared that an answer about what they would do could be used against them. Reasoning that virtually all lawyers would identify themselves as "good lawyers," the author thus felt secure in asking the more general questions about how good lawyers would act.

n222. This table is a composite of T.C. Table L1, at 16-20.

n223. From a societal, moral perspective, it may, of course, be comforting to learn that lawyers would speak out in cases like the hypotheticals. Nevertheless, it is troublesome to find an ethical requirement so extreme that some attorneys would violate it to preserve their own integrity and others would ignore their own consciences in order to abide by it.

n224. Lawyer comments accompanying their survey responses revealed several concrete illustrations of lawyer misconceptions about confidentiality. One criminal prosecutor, for example, seemed to believe that confidentiality rules applied to witnesses. T.C. ID No. L22. Another lawyer, apparently hewing to the traditional line that withdrawal is appropriate in most of the hypotheticals, seemed to believe that disclosure was allowed upon withdrawal:

If an attorney is uncomfortable with a situation, they might consider disassociating themself [sic] from the situation and then consider disclosure, if at all, after severing themself [sic] from the professional relationship.

T.C. ID No. L15 (emphasis added). One real estate lawyer denied that the survey, and thus confidentiality rules, applied to his practice because it did not involve litigation. T.C. ID No. L64.

Lawyers and clients may misunderstand confidentiality rules only to the same extent they misunderstand any complex legal doctrine. The confidentiality context differs, however, in that the primary justification for the rule itself depends on clients' ability to comprehend and rely on the rule.

n225. This table is a composite of T.C. Tables J3A, at 30, 33, 36, 39, 45, 48, 51, 57, 60, 63, 66, 69.

n226. This table is a composite of T.C. Tables J3A, at 31-32, 34-35, 37-38, 40-41, 46-47, 49-50, 52-53, 58-59, 61-62, 64-65, 67-68, 70-71.

n227. That is not to say the respondents would disclose the type of information the hypothetical clients told the lawyers. Respondents may well have assumed they would never be in a similar situation.

This highlights one of the problems in assessing a justification as abstract as "confidentiality encourages clients to reveal information to their attorneys." Of course confidentiality encourages client communications, but how much? As the study suggests, exceptions that cover the information in the hypotheticals might not significantly affect client communications in general. If proponents of confidentiality oppose such exceptions on the narrower ground that confidentiality is needed to "encourage communication of the categories of information in the hypotheticals," a fair evaluation of their argument would weigh the trade-off with respect to each exception. To date, the proponents have resisted such an assessment.

n228. Cf. Burt, supra note 14, at 1018-19 (rules requiring disclosure might "enhance mutual trust between attorneys and clients generally, even though such rules would require attorneys to forfeit trust with some specific clients by disclosing their communications to third parties").

n229. See generally N. CHANNELS, SOCIAL SCIENCE METHODS IN THE LEGAL PROCESS 53-81, 148-81 (1985) (discussing alternative methods of data collection). Survey and opinion research is among the least reliable of scientific evidence. It is largely a hit or miss enterprise. Thus, the Tompkins County study was designed as a tentative first step, an attempt, in part, to determine how a more comprehensive and scientifically significant study might be conducted. In retrospect, the study teaches how many of its questions could have been refined to obtain better and more pertinent information.

n230. For example, 87.5% of lawyers who thought they would get enough information to represent clients effectively without telling the clients about confidentiality were against limiting confidentiality rules. T.C. Table L11, at 22. Of lawyers who admitted they almost never tell clients about confidentiality, 100% were against changes in the rules. T.C. Table L7, at 31. Only a slightly lower figure applies to lawyers who tell clients of confidentiality in less than half their cases. T.C. Table L6, at 18 (89.2%). These results suggest that lawyers may favor confidentiality for reasons other than its effect on clients, perhaps even protecting the guild.

n231. T.C. ID No. L33.

n232. See supra text accompanying notes 191-95.

n233. Ideally, one would of course like to survey clients and the particular lawyers who represented them. Practical impediments, however, including the danger of undoing the attorney-client privilege, may well prevent such a concerted study.

n234. Copies of the questionnaire and results are on file with the author and with the Iowa Law Review.

n235. See supra text accompanying notes 157-58.

n236. Similarly, it seems obvious on the surface that an exception allowing lawyers to disclose planned misconduct would prevent clients from discussing their plans. It would, however, be useful for investigators to confirm this intuition. For clients' desire for advice might cause them to confide in their lawyers nonetheless. The Tompkins County study provides tangential support for that proposition; few clients stated that exceptions covering the information in the Appendix's hypotheticals would cause them to consult with lawyers less. See supra text accompanying note 227.

n237. See supra text accompanying note 145.

n238. See supra text accompanying note 173.

n239. See, e.g., ALA. RULES OF COURT DR 7-102(B)(1); VA. RULES OF SUPREME COURT DR 4-101(C)(3).

n240. See, e.g., RULES REGULATING THE FLORIDA BAR Rule 4-1.6(b)(2).

n241. See, e.g., CONN. RULES OF PROFESSIONAL CONDUCT Rule 1.6(c)(2); MD. RULES OF PROFESSIONAL CONDUCT Rule 1.6(b)(2); NEV. SUPREME COURT RULES Rule 156(3)(a).

n242. The lawyers may, for example, have favored disclosure in particular hypotheticals because they misunderstood the prevailing rules. Or, they may have misinterpreted or added to the facts in a way that would bring the scenario within an existing exception (e.g., future crimes). Similarly, a variety of moral considerations may have prompted different lay subjects to recommend disclosure in any particular case.

n243. See Appendix, hypotheticals (1), (2), (3) & (6).

n244. The statistics cited in the following pages are listed in Tables III, IV, and V supra at 392, 394, and 395.

n245. See Appendix, hypotheticals (4) & (5).

n246. T.C. Table J3A, at 39, 40 (85.5% and 51.7%).

n247. T.C. Table J3A, at 60, 61.

n248. T.C. Table L1, at 17, 19 (24% and 20.8%).

n249. T.C. Table J3A, at 61, 58 (80.3% and 77.4%).

n250. See supra text accompanying notes 203-16.

n251. T.C. Table J3A, at 49, 52, 58 (77.8%, 66.7%, and 77.4%, responding to hypotheticals (6), (7), and (8) respectively).

n252. Only 12.5% and 10.6% of the lawyers believed disclosure was appropriate in hypotheticals (7) and (8). T.C. Table L1, at 18, 19.

n253. 3.8% thought the good lawyer would disclose in the President Nixon hypothetical, 26.9% in the vice president hypothetical, and 13.7% in the Nazi situation. T.C. Table L1, at 20, 19.

n254. The Nazi example may have misled the lay subjects. The client has kept a secret about himself, but has done nothing wrong. The lawyer subjects, in rejecting disclosure, presumably were guided by our legal tradition that unpopular citizens deserve representation. One wonders whether the lay respondents focused on the distasteful identity of the client or the fact that the client concealed the identity from the lawyer. Would they have endorsed disclosure had the client been more sympathetic -- for example, a leftist organization during the McCarthy period?

n255. Over half stated that they would confide less if lawyers gave only vague assurances that they would usually keep information confidential. T.C. Table J1A, at 21 (52.5%).

n256. T.C. Table J1A, at 23 (28.8% of 52.5%).

n257. The exceptions in the model codes are largely permissive. In contrast, the codes in at least a dozen states include mandatory disclosure provisions. Arizona, Connecticut, Florida, Illinois, Nevada, New Jersey, Wisconsin, and Virginia require disclosure to prevent some dangerous crimes or harm to third parties. See I-IV NATIONAL REPORTER ON LEGAL ETHICS AND PROFESSIONAL RESPONSIBILITY (D. Luban ed. 1983). Hawaii, Kansas, Michigan, Ohio, Oklahoma, Rhode Island, South Carolina, and Texas require disclosure of even privileged information to prevent fraud on a person or tribunal. See id.

n258. See supra text accompanying notes 168-69.

n259. Pepper, supra note 24, at 630-32; Shaffer, supra note 53, at 244-50; see also Spiegel, supra note 53, at 118 (requiring lawyers to discuss strategy with clients and obtain informed consent may produce more ethical behavior).

n260. At least some clients are wary of lawyer discretion. When asked how they would react if "the law specifically allowed a lawyer to disclose information received from clients whenever the lawyer 'reasonably deems it to be necessary in the public interest,'" 30.8% stated that they would "trust lawyers less." T.C. Table J1A, at 89.

n261. One abstraction in confidentiality's rationales is discussed supra note 99. Other examples abound. At what kinds of client "dignity" do the rules aim? Although client "trust" is important, do proponents of confidentiality wish to produce trust based on lawyer deception about their trustworthiness? When proponents of rules fail to identify the precise nature of their claims, it becomes very difficult to assess the validity or the effectiveness of the rules -- normatively or empirically.

n262. See Leubsdorf, supra note 16, at 1031 (discussing "market model" of professional responsibility, under which lawyers' duties "should be as negotiable as their price").

n263. As, for example, under an agency model.

n264. Cf. Sirkin v. Fourteenth St. Store, 124 App. Div. 384, 389, 108 N.Y.S. 830, 834 (1908) ("nothing [is] more effective in stopping the . . . spread of [commercial bribery] than a decision that the courts will refuse their aid to a guilty vendor or vendee"); Livingston v. Page, 74 Vt. 356, 361, 52 A. 965, 966 (1902) (candidate's promise to pay newspaper for political support invalid); see also E. FARNSWORTH, CONTRACTS § 5.1, at 329 (1982) ("an attempt to induce conduct may offend public policy though the conduct itself does not, as when a promise is made in return for voting in a particular way").

n265. In the context of administrative law, courts usually must affirm an agency's exercise of discretion, but only if the agency has considered the applicable criteria or standards. R. PIERCE, S. SHAPIRO & P. VERKUIL, ADMINISTRATIVE LAW AND PROCESS § 7.5.1, at 381 (1985); see S.E.C. v. Chenery Corp., 318 U.S. 80, 94 (1943); Mobil Oil Corp. v. Department of Energy, 610 F.2d 796, 801 (Temp. Emer. Ct. App. 1979), cert. denied, 446 U.S. 937 (1980). Appellate courts similarly have reversed decisions because of trial judges' failure to exercise discretion, even though the judges could have reached the same result by considering the appropriate criteria. See, e.g., People v. Massie, 428 P.2d 869, 881, 59 Cal. Rptr. 733, 745 (1967); Capps v. Lynch, 253 N.C. 18, 22, 116 S.E.2d 137, 141 (1960); Paschong v. Hollenbeck, 13 Wis. 2d 415, 422, 108 N.W.2d 668, 672 (1961).

n266. See supra text accompanying note 92; cf. Simon, supra note 74, at 1083 (arguing that "lawyers should have ethical discretion to refuse to assist in the pursuit of legally permissible courses of action").

n267. Arguably, the exercise of moral self examination enhances lawyers' integrity, their public image, and society's trust in the profession. See supra text accompanying notes 96-114.

n268. See Gordon, supra note 24, at 36-38 (discussing different factors that may affect attorneys' willingness to assert independence).

n269. Cf. Burt, supra note 14, at 1032 (current confidentiality rules create no "formal incentive structure . . . to motivate unscrupulous lawyers and clients to seek out one another").

n270. Professor Simon correctly points out that even "selfish" lawyers may place a premium on ethical concerns. Following ethical norms confers "prestige on those identified with them and degradation on those perceived to violate them. Stepping into a role thus degraded has costs that deter people with better opportunities and abilities from doing so." Simon, supra note 74, at 1130; see also Gordon, supra note 24, at 34 (noting influence of informal conventions of propriety upon lawyer conduct); Leubsdorf, supra note 16, at 1032 (lawyers may refrain from inappropriate conduct to secure confidence of clients and colleagues). Moreover, while waiving disclosure may attract one breed of clients, it may deter another.

n271. See MODEL RULES Rule 1.6(a) and CPR DR 4-101(C)(1) (authorizing attorneys to request client approval for unorthodox attorney-client practices). The comments to Model Rule 1.2 may forbid such an arrangement as not "accord[ing] with the Rules of Professional Conduct." But the comments seem to limit their caveat and specifically note the lawyer's right to "exclude . . . means [of representing the client] that the lawyer regards as repugnant." MODEL RULES Rule 1.2 comment 4; see also Rauh, supra note 78, at 17-18 (urging attorneys to negotiate with their clients for conditions that allow more moral conduct in the public interest).

n272. From a purely economic standpoint, if attorneys and clients can pursue their inclinations on an individual basis, codes should not need to include confidentiality rules or exceptions. "Contracting out" should lead to the same results whatever the rules. See generally Demsetz, Wealth Distribution and the Ownership of Rights, 1 J. LEGAL STUD. 223 (1972) (discussing Coase theorem); cf. Schwab, A Coasian Experiment on Contract Presumptions, 17 J. LEGAL. STUD. 237 (1988) (questioning standard view of Coase theorem). In practice, however, strict confidentiality rules provide lawyers with little incentive to bargain for exceptions. Such negotiations may drive prospective clients to attorneys who place less of a premium on matters of conscience. Arguably, the lawyer's "alternative in the real world is not to be more moral but to be amoral or unemployed." G. HAZARD, JR., supra note 104, at 146.

n273. Cf. Gordon, supra note 24, at 34 (first requisite for attorney independence from client desires is "that the lawyer have some source of norms, rules, or conventions to refer to in resisting client pressures"); Simon, supra note 266, at 1130 (some clients consider themselves bound by lawyer conduct embodying important legal "ideals").

n274. Indeed, in-depth discussion of what the lawyer is or is not willing to do may prompt clients to consider alternative representation when they otherwise would not.

n275. The happenstance that educated clients can make best use of their representatives is evident in many aspects of the lawyer-client relationship. Any ethical provision that leaves an aspect of lawyer-client relations unspoken effectively rewards informed clients. See, e.g., MODEL RULES Rule 7.4 and CPR DR 2-105 (limiting lawyers' description of specialties); MODEL RULES Rule 1.16 and CPR DR 2-110 (grounds for lawyer withdrawal); MODEL RULES Rules 3.3-3.4 and CPR DR 7-102 (prescribing limits on attorney zeal); CPR EC 2-19 (suggesting but not requiring advance fee agreement); see also Leubsdorf, supra note 16, at 1031 (courts have "permitted sophisticated clients to waive their rights to require their lawyers to decline representation of other clients who might pose a potential conflict") and authorities at 1031 n. 76.

n276. Even England, our closest counterpart, recognizes public interest exceptions for situations such as hypothetical (2), in which the lawyer learns of information that would exonerate a criminal defendant. See Levine, supra note 62, at 812, 824, and authorities cited at 812 n. 136.

Germany in one sense protects confidentiality to a greater extent than American jurisdictions. It provides not only ethical restrictions, but also civil and criminal liability for a lawyer who improperly breaches confidentiality. However, the German commentators recognize that deciding if confidentiality may be breached in an individual case depends on a balancing of values. See Luban, supra note 39, at 271. The criminal law recognizes that attorneys may speak out "in the case of a serious threat to body, life or freedom." Id. But see id. at 280-87 (questioning when, if ever, German lawyers would in fact disclose).

Thai lawyers apparently do not, in practice, feel bound to preserve client secrets. Although confidentiality rules exist, see Note, Lawyers' Ethics in Thailand, 5 L.J. OF MARUT BUNNAG INT'L L. OFF. 90 (1976), "a client rarely places his confidence in a lawyer except where he has no choice or a large and prestigious firm is involved; and even then, he may tend to trust the 'name lawyers' rather than the firm itself." Nash & Valaisathien, Thailand: The Courts and the Legal Profession, 5 LAWASIA 61, 94 (1974).

Several commonwealth countries, such as New Zealand and Australia, have construed their future crimes and fraud exceptions to attorney client privilege rules broadly. See Peiris, Legal Professional Privilege in Commonwealth Law, 31 INT'L & COMP. L.Q. 609, 637-38 (1982). The European Court of Justice has narrowed attorney-client privilege rules applicable to the European Economic Community by ruling that the privilege applies only to independent, not in-house, counsel. Australian Mining and Smelting Europe Ltd. v. Commission of the European Communities, Case 155/79 E.C.R. 1575, 2 Common Mkt. L.R. 264, 324 P27 (1982); see Stewart & Vaughan, Does Legal Professional Privilege Exist in the EEC?, 72 L. SOC. GAZETTE 1111, 1111 (1975); see also Pagone, Legal Professional Privilege in the European Communities: The AM&S Case and Australian Law, 33 INT'L & COMP. L.Q. 663, 679 (1984) (noting that the High Court of Australia is considering a similar ruling regarding in-house counsel).

n277. The contrary "doomsday" argument parallels the initial reaction to modern civil discovery procedures. There, too, opponents of change argued that liberal disclosure requirements would undermine the adversary system. See, e.g., W. GLASER, PRETRIAL DISCOVERY AND THE ADVERSARY SYSTEM 12 (1968) ("critics have feared that discovery would diminish the advantages of the adversary system"); Sunderland, Scope and Method of Discovery Before Trial, 42 YALE L.J. 863, 867 (1983) ("Every change in procedure by which the disclosure of the truth has been made easier has raised the spectre of perjury to frighten the profession."). Perhaps history teaches that we should not worry unduly about such objections to reform.

n278. In Subin, supra note 26, at 1103-04, Professor Subin discusses the even more dramatic situation in which the kidnap victim may die if the lawyer does not act promptly.

Numerous interesting "disclosure" situations beyond this Article's scope arise when a lawyer represents a client who is or may become a criminal defendant. See generally M. FREEDMAN, supra note 24; Wolfram, supra note 24. At one time or another, most criminal lawyers must consider whether to tell the authorities about (1) the whereabouts of a client who has skipped bail, (2) the possibility the client has committed perjury, see Nix v. Whiteside, 475 U.S. 157 (1986) and authorities cited therein, and (3) client admissions or information that might lead to the discovery of additional evidence or crimes.

n279. Cases have arisen in which lawyers learn from a criminal client that he or she, rather than the accused, has committed a crime. The courts uniformly have held that these communications are privileged and confidential. See, e.g., State v. Macumber, 112 Ariz. 569, 571, 544 P.2d 1084, 1086 (1976) (en banc); State v. Valdez, 95 N.M. 70, 73, 618 P.2d 1234, 1237 (1980). Other societies have created exceptions to confidentiality rules in order to allow disclosure in this context. See A. CROSS, EVIDENCE 398-400 (6th ed. 1985) (discussing English exception). Cf. R. v. Barton, [1972] 2 All E.R. 1192, 1194 (Crown Ct.) (English attorney-client privilege does not apply when documents in possession of solicitor can "help to further the defense of an accused man").

n280. Spaulding v. Zimmerman, 263 Minn. 346, 116 N.W.2d 704 (1962), discussed in G. HAZARD, JR. & W. HODES, supra note 105, at 114; Luban, The Adversary System Excuse in D. LUBAN, supra note 25, at 83, 115.

n281. The position of the CPR on various manifestations of this hypothetical is discussed in Ferren, The Corporate Lawyer's Obligation to the Public Interest, 33 BUS. LAW. 1253 (1978).

n282. See Tarasoff v. Regents of Univ. of Cal., 17 Cal.3d 425, 441-42, 551 P.2d 334, 347 (1976) (psychiatrist has duty in tort to protect patient's threatened victim); see also Hawkins v. King County, 24 Wash. App. 338, 343, 602 P.2d 361, 365 (1979) (lawyer may have tort duty to warn potential victim). In the absence of a direct statement by a client that the client plans to commit a future crime, the CPR seems to require lawyers to remain silent. CPR DR 4-101(C)(3). The Model Rules limit the exception even further to future crimes "that the lawyer believes is likely to result in imminent death or substantial bodily harm." MODEL RULES Rule 1.6(b)(1).

n283. Callan and David discuss a variant of this hypothetical in Callan & David, supra note 26, at 388-89. In a similar securities context, the Securities Exchange Commission has taken the position that attorneys aware of proxy misinformation must disclose. See S.E.C. v. National Student Mktg. Corp., 457 F. Supp. 682, 713 (D.D.C. 1978). The S.E.C. rule has fostered much debate. See generally Gross, Attorneys and Their Corporate Clients: SEC Rule 2(e) and the Georgetown "Whistle Blowing" Proposal, 3 CORP. L. REV. 197 (1980); Gruenbaum, Clients' Frauds and Their Lawyers' Obligations: A Response to Professor Kramer, 68 GEO. L.J. 191 (1979); Hoffman, On Learning of a Corporate Client's Crime or Fraud -- The Lawyer's Dilemma, 33 BUS. LAW. 1389 (1978); Kramer, supra note 63; Lipman, The SEC's Reluctant Police Force: A New Role for Lawyers, 49 N.Y.U. L. REV. 437 (1974).

Arguably, this hypothetical presents a situation in which a lwyer might view the client's fraud as an ongoing "future crime" that justifies disclosure. The comments to the Model Rules, however, make clear that an attorney may not disclose wrongdoing that has already occurred, though he must avoid helping the client to further the misconduct in the future. See MODEL RULED Rule 1.6 comment; see also J. BURKOFF, supra note 18, § 6.5(e) at 6 87.

n284. Arguably, this scenario too involves an ongoing fraud. See supra note 283.

n285. See Agnew v. State, 51 Md. App. 614, 649-56, 446 A.2d 425, 44-54 (1982).

n286. See generally D. LUBAN, supra note 25 (discussing the Ford Pinto case).

n287. Under the CPR, disclosure might be allowed in this situation under the "future crimes" exception. Under the Model Rules, however, nonviolent crimes are no longer covered. Cf. G. HAZARD, JR. & W. HODES, supra note 105, at 117 (discussing similar example).

n288. Callan & David, supra note 26, at 386-88 conclude that the result under CPR DR 7-102(B) is unclear. But see ABA Comm. on Professional Ethics, Opinions, Formal Op. 287 (1953) (disclosure in similar situation forbidden). Under Model Rule 3.3(a)(2), disclosure may be appropriate. But arguably Rule 3.3(a)(2) does not apply either because the lawyer is not involved in the proceeding or because Rule 1.6(b) suggests that disclosure of the client's continuing crime (i.e., perjury) is not permitted because of the nonviolent nature of the crime.