Printer Friendly

Group norms, gossip, and blackmail.

INTRODUCTION I. THE EFFECT OF BLACKMAIL ON GROUP NORMS

A. Blackmail Ambiguously Affects the Expected Cost of Norm Violations

B. Blackmail Impedes the Internalization of Norms

C. Blackmail Inhibits the Communication Processes

That Refine and Reform Inefficient Norms

1. Blackmail Inhibits Gossip-Based

Articulation Of Norm Boundaries

2. Blackmail Inhibits Criticism of

Dysfunctional Norms

D. Norms Theory and Blackmail: Using Law to Create Background Conditions Favorable or Unfavorable to Norms II. THE EFFICIENCY OF THE BLACKMAIL BAN WITH

PRIVACY NORMS

A. The Continuing Puzzle of Opportunistic Blackmail

B. Choosing the Second-Best: Unavoidable Inefficiencies

in the Distribution of Information

1. The Love of Gossip: Why a Blackmail Ban

Produces Excessive Disclosure

2. Asymmetric Transaction Costs: Why a

Lawful Blackmail Market Would Produce

Excessive Secrecy

C. Informational Norms: Why Privacy Norms Correct

the Inefficiency of the Blackmail Ban More

Effectively Than Disclosure Norms Would Correct

the Inefficiency of a Lawful Blackmail Market

1. The Existence of Privacy Norms

2. Privacy Norms Reduce the Excessive

Disclosure Caused By a Blackmail Ban

3. Why Disclosure Norms Would Not Equally

Correct the Tendency of a Lawful Blackmail

Market to Produce Excessive Secrecy

D. Speculating on the Overall Informational

Efficiency of the Blackmail Ban CONCLUSION

INTRODUCTION

Rational choice analysis of law and norms has, to date, yielded several important insights.(1) One might be termed the "substitution hypothesis": legal and norm-based rules are alternative means of social control, and game theory reveals the conditions under which one mechanism governs behavior to the exclusion of the other.(2) That law and norms are alternatives is an old point.(3) What rational choice adds to this insight is some detail about the strategic problems social control mechanisms solve, the incentives individuals face for complying with law or norms, and therefore, a basis for comparing the likely influence of law or norms in a given situation. Robert Ellickson's work, for example, demonstrates the analysis in the context of disputes between neighbors: because property norms solve certain collective action problems, legal rules are not necessary to solve the same problems.(4)

Other rational choice/norms scholars assert what might be termed the "norm governance. hypothesis: Legal rules can shape norms, and the state should cautiously use legal rules to facilitate or obstruct specific norms.(5) Thus, Robert Cooter advocates, in certain circumstances, directly incorporating the otherwise underenforced norms of an industry into legal rules governing contractual relations in that industry.(6) Conversely, certain legal rules may efficiently interfere with dysfunctional norms, as one may regard antidiscrimination laws as usefully impeding enforcement of discriminatory norms.(7) But legal rules may also inefficiently interfere with norms, if for example, the law fails to give adequate weight to informal dispute resolution systems.(8)

The broad purpose of this Article is to illustrate more complex connections between law and norms by complicating the norm governance and substitution hypotheses. My focus is the intersection between group norms and the law of informational blackmail.(9) This intersection is important to both hypotheses. First, the prohibition on blackmail illustrates a more universal type of norm governance. Rather than use law to shape a particular norm, as contract doctrine implements industrial norms or Title VII obstructs discrimination norms, the blackmail ban affects norms "across the board." The state's decision to permit or prohibit blackmail turns out to be an important background determinant of the strength and content of the group norms in society. In this sense, some norm governance is inevitable; even if the state chooses not to shape particular norms, its decision concerning certain behavior, such as blackmail, affects norms generally. The blackmail ban also illustrates a more complex substitution hypothesis. Instead of choosing between a legal and a norm-based rule, the blackmail ban is a case where society has chosen between two bundles of legal and norm-based rules. The efficiency of our current regime depends on how the existing bundle--a legal ban on blackmail combined with certain informational norms--compares to the alternative bundle--a legal right to blackmail combined with the different informational norms that would arise in such a setting.

More narrowly, this Article advances two claims, one positive and one normative. Part I advances the positive claim that blackmail affects the enforcement and content of group norms. Blackmail has certain offsetting effects on the expected cost of violating norms, impedes the internalization of norms, and undermines the communication processes that refine and reform norms. Part II offers the normative claim: that consideration of norms reveals a new justification for the prohibition of blackmail. Here, I first assume the validity of an existing economic theory of the blackmail ban, subject to a particular criticism concerning a subset of blackmail that has been termed "opportunistic."' Part II offers an explanation for why the ban on opportunistic blackmail is efficient, thus supplementing existing theory. My claim is that the regulation of nonproprietary information is necessarily "second-best," the choice being between a regime banning blackmail and producing excessive disclosure and a regime permitting blackmail and producing excessive secrecy. I argue that informational norms obligating privacy correct the former problem to a greater degree than informational norms obligating disclosure would correct the latter.

Before beginning, however, I should make clear the sense in which I am using certain key terms. I follow many others in using "group norms" to mean customary patterns of behavior that individuals within a group feel obligated to follow.(11) And by "group," I have in mind Ellickson's "close-knit" group: "a social network whose members have credible and reciprocal prospects for the application of power against one another and a good supply of information on past and present internal events."(12) Many groups in society qualify as close-knit, and the evidence suggests that many social groups use norms to govern their members' behavior. In his study, Ellickson found that ranchers in Shasta County, California controlled the resolution of property disputes concerning cattle trespass and boundary fences not through law but with norms.(13) The ranchers were connected by both occupation and geography, two elements that are common, though not necessary,(14) to norm-enforcing groups. Steven Cheung, for example, reports that orchard owners in rural Washington provide bees for pollination of their own and neighboring orchards according to a norm requiring a contribution proportionate to the size of one's orchard.(15) Elinor Ostrom describes norms regulating the uses of common pool resources, such as irrigation water, all over the world.(16) Considerable study also shows that norms arise in industries and organizations characterized by repeated interaction among parties.(17)

Finally, throughout this Article I employ the following short-hand: "B" is the person who hers acquired the information useful for blackmail, sometimes a blackmailer; "V" is the victim, usually the person to whom the information pertains; "TP" is the third party (or parties) potentially interested in the information who does not have it, often the recipient of the information should B decide to disclose it.

I. THE EFFECT OF BLACKMAIL ON GROUP NORMS

Most of the existing rational-choice/law-and-norms literature focuses on legal rules of contract or property. But the criminal prohibition on blackmail is one of the few legal rules that directly regulates a mechanism of norm enforcement. The exact definition varies, but many state statutes closely follow the Model Penal Code in defining blackmail (a form of extortion or theft) as existing where one "purposely obtains property of another by threatening to . . . expose any secret tending to subject any person to hatred, contempt or ridicule, or to impair his credit or business repute."(18) Hatred, contempt, ridicule, and impaired reputation are all possible sanctions for violating norms.(19) The meaning of blackmail is, of course, not limited to threatening the shame or reputational loss that occurs for the purpose of enforcing norms. But blackmail includes threats to invoke these norm enforcement mechanisms.

This Part discusses the effect on group norms of prohibiting such threats. I begin the analysis with one simple assumption: that people like to "gossip." People enjoy passing information about people they know to other people they know. Thus, like other forms of communication, gossip is a consumption good, a parstime rather than a burden.(20) Whatever the exact boundaries of this preference, I assume that the fact that another member has violated a group norm is exactly the kind of interesting information people within the group will enjoy passing to others.(21) Given this assumption, and if blackmail threats are deterred by criminal sanctions, we can predict what group members will typically do when they discover that another member has violated a norm: They will disclose the violation to others and, in short order, the violation will become known to the group.(22)

Legalizing and enforcing blackmail contracts, however, will reduce the level of gossip about norm violations.(23) Absent criminal sanctions, blackmail will increase.(24) Thus, rather than gossip, group members who discover norm violations will sometimes conceal their information in order to blackmail the norm violator.(25) Restricting information flow in this manner would affect group norms in three ways. First, legalizing blackmail would change the sanction a group member expects to receive for violating a norm. Second, blackmail would impede the internalization of norms because internalization partly depends on the public punishment of norm violators. Third, blackmail would also inhibit the public communication processes that work to refine norms and to reform dysfunctional norms.

A. Blackmail Ambiguously Aspects the Expected Cost of Norm Violations

A rational group member will violate a norm when the expected benefit exceeds the expected cost. The expected cost of violating a norm depends on the probability of detection and the sanction cost the member will bear if detected. Thus, an individual violates a norm only when b is greater than P(sc), where b is the individual's private benefit from violating the norm, sc is the sanction cost the individual will bear if his violation is detected, and P is the probability of detection by at least one person.(26) The effect of legalizing blackmail is ambiguous because it would likely increase P but decrease sc.

First, for obvious reasons permitting blackmail would increase P. The opportunity for blackmail creates a pecuniary return on investments in information about norm violations. In at least some cases, probably most, the pecuniary return will exceed the value an individual places on gossiping about such information. Thus, individuals will invest more heavily in discovering norm violations,(27) and the probability of detecting a violation will rise.

Second, blackmail would effectively decrease sc, by replacing it with a lower blackmail price. Without a blackmail opportunity, when an individual discovers that a fellow group member has violated a norm, he is likely to circulate that discovery through gossip. A blackmail contract suppresses the gossip that would enable the rest of the group to sanction the violator. A blackmail transaction replaces the sanction cost-sc-with a blackmail price-bp. If bp is equal to sc, then the substitution does not change the expected cost of violating the norm. But I argue that, on average, bp is less than sc. Thus, blackmail lowers the severity of the average sanction imposed on those whose norm violations are discovered.(28)

The first reason V will pay B less than sc is that the two parties will negotiate between sc and B's lower reservation price. Ex ante, there is no reason for V to pay more than sc to avoid incurring sc;(29) therefore, bp can be anywhere between V's and B's reservation prices. B's reservation price is likely to be quite low, much lower than V's reservation price.(30) The outcome of bargaining is partly a function of the differing ability of the parties to perceive each other's reservation price and to misrepresent their own. If skills are randomly distributed, the average outcome may be quite a bit below sc.(31)

Wealth constraints provide a second reason that bp is less than sc. Sometimes V will simply lack the wealth to pay a blackmail price equivalent to sc. For many norm violations the sanctions may be modest, and the likelihood that V cannot pay sc is low. But for norms that may be the most important to the group's welfare--the group's core norms--the sanction may include social and economic ostracism. With these sanctions, it is more likely that V cannot pay sc, and consequently that bp is less than sc. Thus, blackmail will lower the cost of violating the group's most important norms. For these two reasons, the blackmail price will, on average, fall below the value of the norm sanction.(32)

Thus, legalizing blackmail appears to increase the likelihood that someone will detect a norm violation, but appears to decrease the costs the violator incurs once his transgression is detected. The net effect on norm enforcement is indeterminate.(33) For some particular groups or particular norms, it might be possible to speculate productively as to which effect dominates.(34) But I have little basis for guessing whether the blackmail ban, on the whole, increases or decreases the expected cost of violating group norms.(35) The remaining effects, however, are more definitive.

B. Blackmail Impedes the Internalization of Norms

Assume that we cannot predict how blackmail changes the expected cost of violating norms. We still know that the cost the norm violator bears is incurred publicly when it takes the form of a reputational loss, but privately when it takes the form of a blackmail payment. Blackmail drives the punishment of norm violators "underground." In this section and the next, I attempt to predict the consequences of this change. My claim here is that blackmail impedes the internalization of norms.

Many theorists note that norms may arise from the "internalization" of external standards of behavior.(36) To say the norm is internalized is to say that the individual has developed a preference for behaving as the norm requires and suffers some psychic cost--"guilt" or "shame"-from behaving otherwise.(37) Robert Cooter uses signaling theory to construct a rational choice model of internalization. According to Cooter, when a group faces a collective action problem, some of its members recognize and signal to others the collective benefits of cooperation.(38) The resulting "endorsement of cooperation" is "unanimous" because no one gains by defending the strategy of defection; even those intending to defect will publicly endorse cooperation. This unanimous "public consensus" causes people to internalize the norm.(39) Whatever the exact psychology, Cooter's theory is supported by a long line of "conformity" experiments in which individuals facing a unanimous group judgment different from their own tend to conform their expressed judgment to that of the entire group.(40) Sometimes these expressed judgments mask continued disagreement, but sometimes the individuals convince themselves that the group is in fact correct.(41) Cooter suggests that individuals who internalize a norm are willing to bear moderate costs to enforce the norm against others.(42)

Cooter's model, however, appears to claim that the norm arises in a single stage--that for any given individual, the unanimous expression of agreement either causes internalization or it does not. But if unanimity itself fails to produce universal internalization (as Cooter seems to think is probable), then perhaps other mechanisms would come into play. Consider a modification to his model: a "second stage" in which additional internalization occurs. In the first stage, some group members internalize the norm and are willing to bear costs to enforce it. In the second stage, their public norm enforcement causes other members to internalize the norm. A signal of public unanimity combined with individuals bearing costs to enforce the consensus is a stronger signal than public unanimity alone. Strengthening the signal could "convert" to the norm some of those who did not internalize it in the first stage. This effect is particularly likely for children. Public punishment of norm violators is a vivid means of signaling to children the group's view as to the wrongfulness of certain conduct. People are particularly subject to internalization as children, and because children lack the interest or ability to commit many norm violations, they learn at this critical stage by observing the punishment of others.(43) Even for adults, public enforcement may be crucial for inculcating new members with group values.(44)

The publicity of norm enforcement is especially important when considering certain obstacles to internalization. The first obstacle is the "hypocrisy" problem. Cooter's model assumes that in the first stage there is unanimous public signaling, but not unanimous behavioral adherence to that signal. In other words, some people say one thing publicly and do another privately. Perceptions of widespread hypocrisy may impede internalization. People rationalize: Why should I feel guilty when most everyone else does it?(45) But public sanctioning of norm violators permits those who have internalized the norm to demonstrate to those who have not that some group members consistently endorse the norm and bear costs to enforce it. Public enforcement may suppress the contrary hypocrisy signal, causing a new round of internalization to take place.

Consider a second obstacle to norm internalization--a "rationalization" barrier. Like legal rules, some norms are vague. For example, suppose some people internalize a meta-norm that commands them to do "their share" in any group endeavor.(46) On the one hand, such a norm seems to solve all collective action problems perfectly (which suggests that it is too good to be true). On the other hand, the norm is extremely vague. Without a more specific norm, individuals are free to interpret what "their share" is. Here is where rationalization can occur (and can explain why the norm could exist despite free-riding). An individual uses his interpretive creativity to convince himself that his share is as small as possible. He invents rationales for why entirely selfish behavior complies with the norm. Public norm enforcement can undermine this process of rationalization. Public shaming disrupts the comfort of these self-indulgent rationalizations. In public confrontation, a violator can attempt to reduce the sanctions others impose on him either by justifying his violation or exhibiting repentance. But norm enforcers tend to expose mercilessly the self-interested nature of these rationalizations. The violator--and all those who observe the shaming ritual--are thus less likely to rely on those rationalizations again.

Even if these complications are only roughly correct, they sufficiently illustrate how blackmail would interfere with norm internalization. In each instance where B blackmails V concerning his norm violation, there is one less case of public norm enforcement.(47) The fewer instances of public enforcement, the less frequently and intensely the group signals its disapproval of violations.(48) Given that norm violations occur, the weaker signal may ensure that the "hypocrisy" effect will prevent any "stage two" internalizations.(49) Children will witness fewer dramatic examples of the dire consequences of norm violations and have fewer occasions to learn by example. Without public enforcement exposing rationalizations, it will be easier to rationalize selfish conduct as complying with vague norms.

Finally, public norm enforcement may also cause those being punished to internalize the norm. John Braithwaite astutely observes that shaming is a very productive punishment for this reason. Shaming by one's social network may produce guilt mechanisms that control future behavior without external monitoring.(50) But according to Braithwaite's thesis, shaming can produce this effect only if it is followed by "reintegration." Shaming works only to the extent that the shamed individual continues to care about the favorable opinion of his fellow group members. Unrelenting shame pushes the individual, at least psychologically, out of the group; over time, he ceases caring about the opinion that those in the group hold of him. Thus, Braithwaite argues that, except in extreme cases, successful shaming is accompanied by reintegration in which a repentant individual is forgiven and prior relations are resumed.(51)

By rendering impossible the public punishment of a norm violator, blackmail also removes the opportunity for shame and reintegration. Of course, without public shaming, there is no need for reintegration; successful blackmail does not threaten to drive the individual out of the group. The important question, however, is which process is more likely to cause a norm violator to internalize the norm: one where a violator is subject to a shaming ritual, exhibits repentance, and enjoys reintegration; or one where a violator pays a blackmail price to keep his violation secret? I think the question almost answers itself. As stated above, the reason is that public enforcement often exposes the frailty of the norm violator's rationalizations, even to the violator.(52) With blackmail, however, only B shares the norm violator's secret. Yet B has no incentive to puncture the norm violator's rationalizations; he may prefer that V continue violating the norm and subjecting himself to further blackmail. Indeed, B may even lack the "moral authority" to shame V. If B were sufficiently group-regarding, he would disclose the violation rather than seeking to benefit privately from it. The self-serving conduct of the blackmailer may only reinforce the violator's underlying tendency to rationalize his norm violations.(53) In sum, blackmail diminishes public sanctioning, which in turn means observers and recipients of the sanctioning are less likely to internalize the norm.

C. Blackmail Inhibits the Communication Processes That Refine and Reform Inefficient Norms

If norm violators are blackmailed, another consequence is that group members will spend less time in self-conscious discussion of the norm. With fewer cases of public norm enforcement, there is less occasion to discuss whether an act violates the norm, what the appropriate punishment is, or whether the norm requires rethinking. In this Section, I argue that intragroup norm discussion facilitates the efficient development of norms by drawing on two other examples of self-conscious discussion: common law rule-making and consumer complaints. Gossip serves to hone norms just as judicial decisionmaking may refine common law rules and just as consumer complaints may discipline declining firms.

Even those who claim that norms are generally efficient point out the ways in which an inefficient norm can arise.(54) A norm might be inefficient when it first arises; or a norm suited to certain conditions may not change as quickly as those conditions change or may change in the wrong direction. In either case, one may wonder how an inefficient norm can be changed. Self-conscious discussion may refine norms to reflect more precisely the group's interests and reform inefficient norms where they arise.

1. Blackmail Inhibits Gossip-Based Articulation of Norm Boundaries

First, gossip serves as a way of adjudicating disputes over norm violations, and this process, in turn, refines the content of norms to resolve specific concerns.(55) As with legal rules, even if one knows the facts of a situation, there remains the question of applying the norm to the facts. For some norms, the existence of a violation may follow uncontroversially from the facts.(56) But some norms are complex; in addition, it is necessary on occasion to resolve a conflict between norms that may command inconsistent behavior in a particular case.(57)

Through gossip, the group constructs and articulates the exact contours of a norm. Like common law courts, gossip applies the general norm to the particular situation, thereby refining the norm itself.(58) Blackmail would replace gossip and therefore impede this process of refinement.(59) One might object that, whatever the value of adjudication, permitting blackmail would merely reduce, not eliminate, adjudication of norms through gossip. In some cases, observers of a norm violation would choose to gossip despite blackmail opportunities; in other cases, so many people directly observe the violation that there is no opportunity for blackmail. So the question is the marginal value of each gossip-based adjudication. The same difficulty arises in trying to determine the "optimal" level of judicial adjudication: Settlement deprives the courts of an opportunity to state or restate how the law applies to a set of facts,(60) but we can obtain most of the benefits of this judicial refinement without having all cases proceed to trial.

The relationship between gossip and norm refinement is similar: After some point, there are sharply diminishing marginal returns to further discussion. The difficulty is determining whether blackmail would suppress gossip below this point. The comparison to litigation, however, should not obscure a crucial difference: legal rules are formal; norms are not. An important judicial opinion is usually written and published. One opinion may therefore reach a very large audience (at least through lawyers) and continue to reach them for many years. The same is not true of a gossip session. There is no writing, which leaves room for misrecollection and divergent interpretations. Even if there were a transcript, there would be no single authoritative session regarding a particular behavior, nor any single authoritative speaker at a particular session. Thus, it is quite possible that each gossip session has a value for refining and articulating the norm that each adjudication does not have for legal rules. Blackmail could seriously impede this process.

2. Blackmail Inhibits Criticism of Dysfunctional Norms

Beyond refinement, communication may be used to criticize and "repeal" or reform inefficient norms. Criticism may seem to be an unlikely source for making norms efficient. As Albert Hirschman notes in Exit, Voice and Loyalty, economists prefer using the mechanism of "exit" to explain competition.(61) Although exit is crucial, Hirschman explained that in conditions of less-than-perfect competition (for quality), customers will often, at least initially, complain rather than exit.(62) This direct feedback--which he labels "voice"--can discipline market entities and maintain competition.(63)

In the context of norms, "voice" means that group members who are dissatisfied with a norm sometimes complain about it, rather than immediately exit the group. One person may first acquire the information that shows the norm to be inefficient; sharing the information may convince others to abandon or modify the norm. The process may occur early in norm formation, thus preventing the group from fully embracing an inefficient norm, or it may occur later, after changed conditions render an existing norm inefficient and the critic points out that the old norm no longer serves the group's interests.(64)

Blackmail will significantly restrict intragroup criticism of norms. Initially, note that an individual may bear a cost in criticizing norms. Most group members are likely to believe the norm serves the group's interests; some have even internalized the norm. The group may therefore infer that the critic is attempting to undermine a norm because he is covertly violating it.(65) An individual will criticize a norm when he values the increased probability of changing the norm more than the increased probability of being sanctioned.(66) Because the probability that any one member's criticism will change the norm is very low, public criticism of norms is a public good apt to be undersupplied.(67)

With blackmail, an individual deciding whether to criticize a norm will consider two additional factors: how his criticism affects the possibility of paying blackmail or receiving blackmail payments. Consider first the possibility of being blackmailed. Blackmail will change the benefit of norm criticism if it changes the expected cost of violating an inefficient norm. If the legality of blackmail raises (lowers) the expected cost, it would raise (lower) the benefits of reforming the norm. Part I.A demonstrated that blackmail has an ambiguous effect on the expected cost of violating a norm,(68) and thus an ambiguous effect on the benefit of reform. The effect on the costs of criticism, however, is unambiguous. The possibility for blackmail creates an expectation of profits from blackmailing others after discovering their norm violations. The profits create an opportunity cost to reforming inefficient norms. Thus, the only determinate effect of blackmail is to increase the costs of norm criticism, lowering the odds of reforming inefficient norms.(69)

Even if we were to assume, contrary to the claim of Part I.A, that legalizing blackmail would increase the expected cost of violating norms, legalization would still probably inhibit norm reform.(70) By this contrary hypothesis, the net effect of blackmail on criticism would depend on the sum of two offsetting effects: (1) legalization would increase the expected cost of violating norms and therefore increase the benefit of reforming inefficient norms; and (2) legalization would create expected profits from blackmailing norm violators and therefore create costs to reforming even inefficient norms. The latter result would tend to dominate because of how the effects are distributed: legalizing blackmail would disproportionately raise the costs of criticism for those members who are likely to be the most persuasive norm critics.

To understand the point, consider two simple observations about signalling and the "persuasiveness" of norm criticism. First, other things being equal, the perception of sincerity makes criticism more persuasive. Sincerity does not ensure persuasiveness; group members will typically think criticism is mistaken. But the perception of insincerity nearly guarantees unpersuasiveness, so a critical signal that other members believe to be sincere is more likely to persuade them. Second, the perceived sincerity of norm criticism is inversely related to the amount others believe the critic would personally gain from norm reform. Those who gain the most from violating a norm will want it "repealed" and will claim it is inefficient even if they actually believe the norm is beneficial to the group as a whole. If other group members believe a critic will substantially gain from norm repeal, they are therefore more likely to perceive his criticism as being insincere. Conversely, those who stand to gain the least from norm repeal will be the most persuasive critics.(71) Their criticism signals that they actually believe the norm is not beneficial to the group.

Legalizing blackmail would affect the most persuasive critics differently than the least persuasive critics. Both subgroups would gain an equal opportunity for blackmailing others for violating norms; for both, criticizing a norm would risk losing this opportunity. Thus, the marginal costs of norm criticism are the same for each subgroup. But if legalizing blackmail increases the benefits of norm reform (again, contrary to Part I.A), the marginal benefits of criticism would not be equally distributed. Those with the greatest stake in violating a particular norm would lose the most when the legalization of blackmail increases the expected cost of violating the norm; conversely, when the norm is repealed, this subgroup would gain the most by eliminating the enhanced norm enforcement that blackmail made possible. Thus, those who gain the most from violating a norm--the least persuasive norm critics--will accrue most of the new benefits of norm repeal. On the other hand, those who gain the least from violating a norm--the most persuasive critics--will gain a relatively small fraction of the marginal benefits of repeal.

Because the benefits of criticism differ for these two groups while the costs are equal, their willingness to criticize the norm will often differ. The less one gains from violating a norm, the more likely it is that the marginal costs of norm criticism (foregone blackmail profits) will outweigh the marginal benefits of criticism. The converse is true for those who gain the most from violating a norm. Thus, legalizing blackmail will disproportionately suppress criticism by those with the least interest in violating a norm, the criticism other group members are most likely to perceive as being sincere. To illustrate, if legalizing blackmail raised the costs and benefits of norm criticism by exactly the same average amounts, legalization would have no net effect on the quantity of norm criticism, but would still increase the quantity of the least persuasive criticism and decrease the quantity of the most persuasive criticism. Of course, if legalizing blackmail were to raise the costs of violating norms sufficiently, it might raise the benefits of reforming inefficient norms enough to produce a rise in all norm criticism. But given the empirical uncertainties, the structural factors create a greater risk that legalization will suppress persuasive criticism,(72)

Finally, legalizing blackmail would also inhibit some of the less persuasive criticism in the one circumstance where such criticism might matter. Imagine that a group deters virtually all criticism of an inefficient norm by severely sanctioning those who speak against the norm. In this situation, the only source of criticism may be norm violators who speak out after their violations are discovered and publicized. When a member's violation is exposed, he may suffer no additional sanction by initially criticizing the norm or some part of it.(73) Of course, these critics will be highly unpersuasive for the reasons discussed above: others will doubt the sincerity of criticism from those who, facing sanctions, stand to gain the most from undermining the norm. But even if there is only a very low probability that such criticism will spur the group to rethink the norm, in the situation described these members represent the only source of criticism, the only chance for initiating a dialogue critical of the norm. Blackmail, however, would silence this subgroup. When a group member secures the secrecy of his norm violation by paying a blackmail price, he will now suffer an additional public sanction by criticizing the norm.(74) By driving the punishment "underground," blackmail gives the violator a reason to keep his criticism underground as well. Thus, at the point where an inefficient norm is most difficult to reform, blackmail may guarantee its stability.(75)

In sum, blackmail works against "voice,. an important mechanism for refining and reforming inefficient norms.

D. Norms Theory and Blackmail. Using Law to Create Background Conditions Favorable or Unfavorable to Norms

When discussing norm governance, most rational choice scholarship proceeds in the following manner. One identifies a specific norm and then determines whether a legal rule could efficiently strengthen or weaken that norm. Eric Posner has discussed a less direct type of legal regulation of norms.(76) By directly regulating solidary groups (through "group-based rules"), the state can expand or contract the power of such groups to use informal mechanisms of social control. This insight also applies to indirect forms of regulation. As Posner notes, even legal rules with universal application ("category-based rules") can, by allocating entitlements to individuals, affect the returns of cooperating with or defecting from solidary groups.(77)

The blackmail example demonstrates another indirect means by which law can regulate norms. The blackmail ban does not directly shape a particular norm, nor does it operate on norms by first making membership in norm-enforcing groups more or less attractive to individuals. Instead, by making norm enforcement public, the blackmail ban directly regulates the means by which groups enforce, internalize, and reform norms. Yet the ban works only indirectly in that it is not focused to facilitate or impede any particular norm, nor the norms of any particular group. The blackmail ban is a background legal rule determining the power and content of group norms across the board.

The positive analysis of this Part yields certain normative implications. One is that if norms tend to be efficient, as certain commentators claim,(78) then the blackmail ban is efficient because it facilitates the internalization and refinement of norms. And if it is socially desirable that norms maximize the welfare of the group, then the blackmail ban is efficient because it facilitates the communication processes by which a group reforms dysfunctional norms. But if, as other commentators seem to claim,(79) norms are generally inefficient, then this Part reveals a novel benefit of legalizing blackmail.(80) The average efficiency of norms is likely to remain controversial for some time, requiring considerably more theoretical and empirical work to resolve. For this reason, the normative evaluation of the blackmail ban advanced in the next Part does not depend on the general efficiency of norms, but on the efficiency of a particular informational norm.

Even without resolving the average efficiency of norms, however, the blackmail example reveals greater complexity to the state's task in norm governance. The state must select background legal rules in combination with rules directly regulating norms. Thus, in addition to policies that directly support and impede the enforcement of efficient and inefficient norms, respectively, we may want the state to create background conditions either generally favorable or hostile to norms. Indeed, even if it were impossible or imprudent for the state to regulate specific norms, it may not be possible for the state to remain neutral regarding norms as a general matter. Certain legal rules, like the blackmail ban, will have widespread if unintended consequences for norms.

In choosing either to favor or disfavor norms generally, several factors determine the optimal background condition. For example, if most norms are efficient, the state should create background conditions favorable for norms and then enact legal rules interfering with the minority of norms that are inefficient. The converse is true if most norms are inefficient. A second factor is the relative effectiveness of the specific norm-shaping rules. For example, if it were generally easier to use centralized power to impede than to facilitate specific norms, that fact would favor creating background conditions favorable to norms and then using the more cost-effective specific-interference rules for curbing inefficient norms. Again, the converse is true if centralized support of norms is more effective than centralized interference. Whatever the efficient combination of rules, it is important first to identify the legal rules that affect the basic background conditions for norm formation and enforcement. The blackmail ban is one such rule.

II. The Efficiency of the Blackmail Ban with Privacy Norms

In recent years, a great many scholars have attempted to explain precisely what is wrong with blackmail.(81) The prohibition on blackmail presents something of a puzzle: The blackmailer is punished for giving his victim a choice between paying a price and suffering an embarrassing yet lawful disclosure of information. The law prohibits the threat to disclose information even when the disclosure itself is neither criminal nor tortious. If an individual is legally entitled to tell the world some secret fact about a person, why is he forbidden from selling to that person his right to make the disclosure? The choice would seem to make the "victim" better off in any instance in which he chooses to pay the blackmail price.

This Part addresses this much debated normative issue. Consideration of group norms reveals new efficiencies in the prohibition on blackmail. These efficiencies do not by themselves justify prohibiting blackmail. Instead, I begin with an extant economic theory of the ban. But, assuming the reigning theory is otherwise valid, it is underinclusive, justifying only part of the prohibition. My claim is that the existence of certain norms explains why the ban extends as broadly as it does, that an analysis of norms fills a significant gap in existing theory.(82)

Part II.A explains this theoretical gap: the problem of "opportunistic" blackmail, where the blackmailer threatens to disclose information he obtained adventitiously, not as a result of investment.(83) The remainder of the Part then explains why the ban on opportunistic blackmail is efficient. Part II.B argues that the regulation of adventitiously acquired information is necessarily "second-best," the choice being between a regime banning blackmail and producing excessive disclosure (because of a desire to gossip) and a regime permitting blackmail and producing excessive secrecy (because of a market failure due to asymmetric transaction costs). Part II.C then claims that informational norms obligating privacy correct the first problem to a greater degree than informational norms obligating disclosure would correct the second. The conditions in which privacy and disclosure norms arise ensure that the former will be more powerful than the latter. Thus, the relevant social control choice is not between law and norms, but between one bundle of legal and norm-based rules and another. The legal rule banning blackmail is preferable because the norms that arise along with the ban make this bundle more efficient than a legal rule permitting blackmail along with the norms that would exist in tandem with that rule.

A. The Continuing Puzzle of Opportunistic Blackmail

Economic theorists have made considerable progress in explaining the prohibition on blackmail. To summarize briefly a substantial literature--at the risk of omitting important details--the central evil of blackmail is that it induces investment in a wasteful or "sterile" activity.(84) If blackmail were legal, individuals would expend resources acquiring negative information about others in order to extract a payment from them. Viewed ex ante, legalizing blackmail makes victims worse off because it creates an incentive for blackmailers to discover secrets they would otherwise never bother to learn. The victim's payment to the blackmailer is a wealth transfer benefitting the blackmailer exactly as much as it harms the victim. But the blackmailer's initial investment in gaining the blackmail information ensures that his net gain is less than the victim's loss. In short, the industry consumes resources merely to transfer wealth rather than to create it.(85)

The blackmail transaction would not be sterile, however, if it provided useful incentives to potential victims. Much of the literature attempts to evaluate the incentives blackmail creates. The most discussed example is crime: if the state permitted the blackmail of criminals, perhaps it would raise the cost of committing crime.(86) But, as Part I.A explained, blackmail can either increase or decrease the cost of a secret activity. Although legalizing blackmail would increase investment in discovering crimes, it would also cause people to conceal criminal discoveries they would have made without investment. As Steven Shavell notes, some members of this latter group, who would have volunteered information to law enforcement when blackmail was illegal, will now offer the criminal the option of replacing a more severe sanction (such as prison) with a less severe sanction, the blackmail price.(88) Although there are many further wrinkles in the analysis," the net effect is essentially ambiguous and the literature is, to my mind, inconclusive. Incentives do not provide a compelling argument for legalizing some subset of blackmail, but neither do they create an independent argument for prohibition. The economic case against blackmail is not that we know the incentive effects are bad but that we doubt the effects are good and we do not want substantial resources invested in what is probably a sterile activity.(89)

The wasteful investment theory, however, suffers from a substantial weakness. The theory really explains only why we prohibit people from investing in the blackmail enterprise. Jim Lindgren, who first made this objection to the economic analysis of blackmail, states the matter this way: Economic theory explains why we prohibit "commercial" blackmail, that is, where the blackmailer invested in gaining the secret information. But the theory does not explain why we prohibit "opportunistic" blackmail, that is, where the blackmailer adventitiously discovered the secret information.(90) If B does not invest in but merely stumbles upon negative information about V, why is B's blackmail of V criminal?

Lindgren's challenge recalls Anthony Kronman's distinction concerning information disclosure during contract negotiation.(91) Kronman says that contract law should not, and typically does not, require disclosure of information a party acquired through investment, but should and typically does require disclosure of information adventitiously acquired. For Kronman, the point of the distinction is to protect socially useful investments in information. Lindgren makes the same distinction when seeking to suppress socially unproductive investments in information: Prohibiting the use of information one has acquired by investment does not require prohibiting the use of information one has acquired by accident.(92)

To date, there has been no adequate answer to Lindgren's criticism.(93) Some theorists have responded by asserting that adventitious discoverers of information have no reason to bear the costs of publicizing their discovery (except perhaps to establish a reputation useful for future commercial blackmail). Permitting blackmail will encourage B to threaten a disclosure he would not make if blackmail were illegal.(94) The result is to make V worse off and to waste the transaction costs necessary to reach a blackmail agreement. But Lindgren's retort is simple: Why assume that disclosing information is a "cost"? People disclose information all the time-they talk, they write, they publish-and they seem to enjoy it.(95) Even the disclosure of negative information about others-sometimes called "gossip"--is, for many, a favorite pastime, a consumption good rather than a burden. Lindgren is surely right on this point.(96) If so, many adventitious discoverers of negative information will gossip about it, unless paid for their silence. If V is willing to pay B to forgo the pleasures of gossip in this circumstance, why should we prohibit what appears to be a pareto-superior trade?(97) The existence of group norms provides a possible answer.

B. Choosing the Second-Best: Unavoidable Inefficiencies in the Distribution of Information

By definition, permitting opportunistic blackmail does not cause any investment of resources, wasteful or otherwise, into acquiring information for blackmail purposes. The question, then, is whether permitting such blackmail will produce a more efficient distribution of adventitiously discovered information, that is, whether those who value such information the most will receive it. The problem is complex because for some information widespread circulation is efficient; for other information, circulation is inefficient. It furthers group interests for everyone to acquire information, for example, about crimes, torts, norm violations, or the occupational competence of market players (for example, that a dentist in the neighborhood is particularly good at dealing with children). But the group suffers if gossip circulates private information the disclosure of which harms the individual subject more than it benefits participants in the gossip.(98) Thus, the efficient outcome is somewhere between complete disclosure and complete secrecy. Unfortunately, neither the regime banning blackmail nor a regime permitting blackmail is likely to achieve the optimal distribution of information.

1. The Love of Gossip: Why a Blackmail Ban Produces Excessive Disclosure

Assuming perfect enforcement, the blackmail ban prevents the subject of information--the blackmail victim--from paying to preserve its secrecy. Because the blackmail ban does not prevent third parties from paying for the disclosure of information, the ban seems to ensure that more information will be disclosed and disseminated than is efficient. Indeed, because many people who enjoy gossip will pass information for free,(99) banning blackmail appears to ensure nearly universal disclosure. On the other hand, absent some market failure, one should expect that a free market in blackmail, where all parties are permitted to bid for secrecy or disclosure, would achieve the optimal distribution of information. Thus, if one imagines information disclosure as a continuum ranging from "excessive secrecy" to "excessive disclosure" with "optimal disclosure" somewhere in between, one might presume that the market can achieve optimal disclosure and that the blackmail ban ensures excessive disclosure.

Absent consideration of norms, I assume that this analysis is correct in concluding that the ban on blackmail will produce excessive disclosure. But, for the reasons addressed next, the free-market rule permitting blackmail will also fail to achieve optimal disclosure.

2. Asymmetric Transaction Costs: Why a Lawful Blackmail Market Would Produce Excessive Secrecy

Market transactions-including blackmail-appear to be an efficient means of allocating information. People have nonproprietary secrets and some willingness to pay to keep those secrets. After B discovers V's secret, B can negotiate with V for a blackmail (secrecy) contract,(100) or with TP for a disclosure contract. At this point, the transaction costs might be fairly low in comparison to the gains V or TP might make from a blackmail or disclosure contract with B. The situation is analogous to B having a single good for sale, which TP wants, but which V does not want TP to have. An auction in which V and TP bid against each other for the good would seem a suitable way to resolve the matter efficiently. If the market worked in this manner, there would be no instance where B inefficiently conceals V's secret from TP; barring nonmarket behavior (for example, violent threats), the only inefficient secrets would be those known only to the subject of the secret (V).

But a free market in blackmail would actually produce excessive secrecy. When the good is nonproprietary information, the market fails to handle the conflict efficiently.(101) In this three-way negotiation, there are two sets of transaction costs--those for the B-V agreement (for secrecy) and those for the B-TP agreement (for disclosure). Obviously, the outcome depends in part on the size of each transaction cost relative to the corresponding contractual surplus-there being no gains from trade if the former exceeds the latter. But it also matters what the transaction costs are relative to one another. Assume that V bids $x for secrecy, that TP bids $y for disclosure, that the transaction costs for the B-Vtransaction are [tc.sub.B-V], and that the transaction costs for the B-TP transaction are [tC.sub.B-TP]. Assuming at least one value is positive, B will choose the higher of the values (x- [tc.sub.Bv]) and (y - [tc.sub.B-Tp]). An asymmetry in transaction costs will cause B to choose inefficiently when x > y > (y - [tc.sub.B-Tp]) > (x- [tc.sub.B-V]), in which case he will disclose even though V is willing to pay more than TP. Similarly, when y > x > (x - [tc.sub.B-V]) > (y - [tc.sub.B-TP]p), B will remain silent even though TP is willing to pay more than V.

The potential for asymmetry matters because TP faces systematically higher transaction costs than V, and the resulting asymmetry leads B to preserve secrecy excessively. There are at least two factors working to ensure that tcH TP > tcB V. First, transaction costs increase with the number of people involved in the bargaining. In the blackmail scenario, there are likely to be more TPS than Vs. The reason is that B's information is secret. With rare exceptions, all the Vs already know the information that they would prefer to keep secret. Yet the more people who currently have the information, the less likely it is that knowledge of the information is limited and controlled, and therefore the less likely that the information remains "secret." Thus, the fact that information pertaining to large numbers of Vs is less likely to be a secret naturally limits opportunity for blackmail in such cases.(102) But there is no such limitation on the number of TPs--the people who would value the information, but who are unaware of its existence. Thus, it is likely that B will need to bargain with more TPs than Vs. Therefore, on average B will incur higher transaction costs extracting the value of disclosure from TP than extracting the value of secrecy from V.(103)

Second, the transaction costs for the B-TP agreement (disclosure) exceed *'e transaction costs for the B-V agreement (secrecy) because of what might be termed an informational bargaining dilemma. As Thomas Schelling aptly noted: "It is hard to sell a secret without giving it away."(104) It is difficult to extract the value of information from a buyer without describing the information to be sold, yet describing the information may effectively give away what one is trying to sell.(105) The problem arises when the information is not proprietary and one can easily exclude others from using the information only by keeping it secret.

Consider an example: B learns that V is having an adulterous affair. B now attempts to sell that information to TP, V's wife, whom B correctly assumes would place a high value on such information. If B tells TP all that he knows and then asks for payment, TP has no reason to pay for what she already has. If B merely tells TP she should pay him a specific sum after which he will give her something equally valuable, she has no reason to trust him sufficiently to make the payment. Suppose B adopts an intermediate strategy of describing some but not all of the information, stating to TP, for example, that he has information "about your husband" or information "about your marriage." Now he risks both sides of the dilemma: (1) that he has disclosed too much and TP can, with minimal investigation, discern the remaining information, or (2) that he has disclosed too little and TP will not believe he really has anything of value.

There are ways of solving this problem, but few are available to the opportunistic blackmailer, and those that are available significantly raise the transaction costs of B's dealing with Tp.(106) Reputation is one solution, but in the context of opportunistic blackmail B is not likely to have a reputation as a person who always names a fair price for his information.(107) There are also contractual

solutions that work well in other contexts: The parties could agree (1) that TP promises to make no "use" of B's information after he discloses it unless TP first purchases B's assent; or (2) that if B's information "benefits" TP, B will share a percentage of that benefit. But while a contractual solution is effective in situations where TP's use of the information is easy to define and verify, much blackmail information concerns embarrassing facts that are useful merely by knowing of their existence. In other cases, the use of damaging facts may be difficult or impossible to verify. Moreover, much of the benefit TP derives from the information may have no, or no easily measured, economic benefit for B to share.

Note that where B knows a secret of V-the blackmail situation--the law does not prohibit B from selling that secret to TP. Yet despite the legality of such transactions, they seem to be rare. Lindgren, for example, argues that banning opportunistic blackmail causes B to disclose excessively, not because TP will pay B to disclose--but because B will gain the intrinsic pleasure of gossip. The absence of such transactions between B and TP is explained by the informational bargaining dilemma B faces in such situations. On the other hand, the fact that B does occasionally manage to blackmail V, despite the potential criminal sanctions,(108) demonstrates that bargaining with V does not raise the informational bargaining dilemma. V already knows the information, or if he does not know it, he is still subject to blackmail once he does. Bargaining with V does not diminish or risk diminishing the value of B's asset.

In short, asymmetric transaction costs provide V a comparative advantage over TP in bribing B; therefore, an unfettered blackmail market would produce inefficiently low levels of disclosure.(109)

C. Informational Norms: Why Privacy Norms Correct the Inefficiency of the Blackmail Ban More Effectively Than Disclosure Norms Would Correct the Inefficiency of a Lawful Blackmail Market

The ban on opportunistic blackmail produces inefficiently high disclosure because, absent market payments, B will gossip even when his benefit is less than the loss imposed on V. But a free market in opportunistic blackmail produces inefficiently low disclosure because asymmetric transaction costs (ATC) ensure that V outbids TP even when TP values disclosure more than V values secrecy. Thus, we are faced with a policy choice between two flawed options. Determining which inefficiency is greater is quite difficult, apparently leaving no powerful reason for choosing one over the other.(110)

When we introduce the possibility of informational norms, however, we discover a basis for selecting one rule over the other. An asymmetry exists that permits informational norms to correct the inefficiency of the blackmail ban more effectively than they correct the inefficiency of the lawful blackmail regime. Indeed, one reason why it is difficult to determine which suboptimal choice is better, without discussing norms, is the difficulty of predicting what B will do with information in the absence of monetary incentives. Because asymmetric transaction costs block B-TP agreements for disclosure and the blackmail ban blocks B-V agreements for secrecy, B's decision must be made without monetary inducements. I argue that this result is the whole point of the ban-that the absence of such inducements creates optimal conditions for informal norm enforcement. I proceed by (1) identifying the existence of privacy norms, (2) explaining why they partly correct the inefficiency of the blackmail ban, and (3) demonstrating that disclosure norms would not provide an equivalent correction for the inefficiency of a lawful blackmail market.

1. The Existence of Privacy Norms

Social groups face a collective action problem concerning "appropriately private" information-information the disclosure of which harms V more than it benefits TP.(111) Each person may gain some small enjoyment from passing along such titillating information, but the group as a whole is worse off when everyone behaves that way. For example, disclosure of certain medical conditions, physical abnormalities, sexual practices, or victimizations may prove mildly interesting to others, but cause great embarrassment to the subject whose privacy is thereby disturbed. Indeed, the problem is severe because individuals have another incentive to pass information that is detrimental to group welfare. Gossip can benefit the gossiper in the same relative way that other means of "putting down" others can benefit an individual. In contrast to the target of the gossip, the gossiper and gossip recipient are positively distinguished.(112) People are strongly motivated to obtain relative position or social status.(113) One means of gaining relative position is to lower the position of others. If everyone in a social group follows this tactic, however, the result is destructive to the group.(114) Consequently, groups need especially to restrain individuals from using gossip in this strategic manner.(115)

My claim is that groups routinely enforce informational norms--"privacy norms"--that work to maximize group welfare by reducing inefficient disclosures of appropriately private information. The empirical claim is that there are norms requiring that one not seek certain kinds of information absent compelling reasons and also that one not disclose certain kinds of information except to certain persons for certain reasons.(116) Some of the rules limiting acquisition of information are sufficiently common to be genuinely "social," as opposed to merely group, norms; for example, absent unusual circumstances, one must avoid reading another person's mail or diary, peering into the windows of another person's home, or intruding upon the seclusion of another person's shower or toilet. More generally, in social settings, one develops a reputation for rudeness by failing to accommodate others who wish to hold a private conversation or by commonly asking for information that people prefer not to disclose.(117)

Nondisclosure norms are more complex because they are more sensitive to reasons and relationships. As Ferdinand Schoeman notes, if one knows a friend has been raped, one might, without violating a privacy norm, tell other friends of the victim who would provide her emotional support.(118) But if a person discovers that a mere acquaintance has been raped, disclosing that fact to every other acquaintance of the victim would incur considerable social disapproval.(119) The same is true of many other facts. Privacy norms do not forbid everyone in every circumstance from disclosing the fact that a person was adopted, suffered incest as a child, had a spouse who committed suicide, accidentally killed someone, enjoys a particular sexual activity, or suffers from cancer, AIDS, sterility, or impotence. For each example, one can imagine cases where limited disclosure incurs no sanction, such as where one discloses the information to a psychiatrist or doctor who provides treatment for the subject, to a close friend who will provide comfort to the subject, or perhaps to an individual seeking a relationship with the subject in which one of these facts would be relevant. But disclosing such information merely for its amusement value, to demonstrate that one is "in the know," or, worse, for the purpose of inflicting harm, will commonly incur serious social sanctions. Even disclosing for a sufficient reason will incur some disapproval if one discloses the private information more widely than necessary.

The examples also demonstrate my normative claim that privacy norms tend to maximize the welfare of the group in which they arise. Where the harm of disclosure is manifest, these norms are sensitive to the benefit of disclosure and are more likely to permit disclosure where the benefit is great.(120) On the other hand, privacy norms most emphatically condemn disclosure generated by the group's collective action problem, where disclosure generates only mild amusement or relative distinction. Consider the apparently conflicting norms concerning gossip. As many scholars have noted, in a wide variety of societies and social groups, "gossip" is both widely practiced and widely condemned.(121) I do not believe the explanation is hypocrisy. Instead, we often practice gossip either by passing essentially harmless information for its amusement value or by passing genuinely harmful information for adequate reasons. But when we condemn "gossip," we use the term narrowly to refer to cases where the gossiper passes damaging information without adequate reason,(122) and particularly where the gossiper is acting strategically to advance his relative interests by sabotaging others. That this latter form of gossip is so widely condemned is evidence of the strength and pervasiveness of privacy norms.

Definitive proof of these claims awaits a close analysis of the informational practices of particular groups. I suspect that close-knit groups enforce privacy norms more tailored to their circumstances than the above examples suggest. For instance, the residents of Cavendish, Vermont apparently refused to divulge the location of Alexander Solzhenitsyn's home to most inquirers for the seventeen years that he lived there.(123) But the anecdotal evidence of privacy norms is sufficient to make my descriptive and normative claims plausible and to merit an investigation into the effect of such norms on the efficiency of blackmail.

2. Privacy Norms Reduce the Excessive Disclosure Caused By a Blackmail Ban

Absent norms, the blackmail ban creates the inefficiency of excessive disclosure. But privacy norms work to correct this inefficiency. Each privacy norm carves out an exception to the baseline of gossip. Upon discovery of information about V, B will now gossip in some cases and, respecting privacy norms, remain silent in others. If the privacy exceptions tend to be welfare maximizing-if they arise where disclosure does more harm than good-then B will remain silent in cases where silence promotes group welfare. Thus, we do not have to choose between complete secrecy (where blackmail is permitted) and complete disclosure (where blackmail is banned). By removing all the monetary incentives from B's decision--by creating a space of market inalienability--we maximize the power of norms. Those norms can then prevent some of the most inefficient disclosures while the love of gossip ensures that the ATC market failure does not generally prevent disclosure.

Of course, one might ask why the norms situation cannot be reversed: Permit blackmail but hope that a norm of disclosure arises for situations where secrecy is inefficient. The next Section considers that possibility.

3. Why Disclosure Norms Would Not Equally Correct the Tendency of a Lawful Blackmail Market to Produce Excessive Secrecy

If blackmail were legal, norms of disclosure might arise to ameliorate the problem of excessive secrecy. But disclosure norms would be less successful at correcting the ATC market failure than privacy norms are at correcting the inefficiency of the blackmail ban. The primary reason is that disclosure norms would likely be weaker than privacy norms. In addition, the state can supplement norms of privacy or disclosure with direct regulations concerning the dissemination of information, but laws requiring disclosure would probably be less effective than laws requiring privacy.

First, recall that an individual violates a norm when b is greater than P(sc), meaning when the benefit of the violation exceeds the probability of discovery times the sanction cost one incurs upon discovery.(124) Thus, the more an individual benefits from a behavior (an increase in b) and the more easily the behavior is concealed (a decrease in P), the more difficult it is to enforce a norm against the behavior. For both reasons, privacy norms will be more powerful than disclosure norms.(125)

On average, the benefit of violating a disclosure norm will be greater than the benefit of violating a privacy norm. The reason is the asymmetric transaction costs discussed above: V bears lower transaction costs in dealing with B than TP does. Thus, V will be able to offer B higher bribes for violating disclosure norms than TP will be able to offer for violating privacy norms. In other words, if blackmail is permitted, V will frequently be willing to pay a substantial sum to B to induce his violation of a disclosure norm. This bribe pits B's norm compliance against his tangible financial gain. But with a blackmail ban, high transaction costs work against TP's bribing B to violate privacy norms. Because TPs typically do not know that the information they seek exists, they face prohibitive transaction costs in attempting to bribe B to disclose his discovery. The transaction costs' asymmetry creates an asymmetry in the benefits of violating norms of privacy and disclosure.

Second, on average, the probability of discovery is greater for violations of privacy norms than for violations of disclosure norms. The reason is simply that violations of privacy norms involve improper disclosure while violations of disclosure norms involve improper nondisclosure. Disclosure is inherently more public--more likely to be detected--than nondisclosure. Admittedly, people may try to disclose anonymously when norms demand silence, but disclosure of appropriately private information necessarily alerts group members of the existence of a norm violation. At that point, what remains unknown is the identity of the violator. However difficult it may be to identify that person, it is not as difficult as the detection of disclosure norm violations because nondisclosure does not necessarily alert anyone of its existence. Instead, norms obligating disclosure face the unique difficulty that their violation tends to be a complete secret. If B is the only one who knows of V"s substantive norm violation, the optimal condition for blackmail, then B faces a very low risk of sanction should he fail to disclose. Thus, the probability of punishment is higher for violating privacy norms than it is for violating disclosure norms.

In sum, privacy norms command more compliance than disclosure norms.(126) The transaction costs' asymmetry creates an asymmetry in the benefits of violating norms of privacy and disclosure. And the unique difficulty of detecting when someone withholds a secret creates an assymmetry in the possibility that a violation of these different norms will be discovered. Because the benefits from violating disclosure norms are higher and the risk of detection is lower, B will more frequently violate such norms.(127) Thus, the current regime's bundle of legal and norm-based rules exploits the relative strength of privacy norms. The legal rule banning blackmail is desirable because it works in tandem with norms that will command greater compliance than the norms that arise with a regime of lawful blackmail.

There is a final point that favors privacy norms over disclosure norms. The state can more effectively supplement privacy norms by direct regulation. In theory, one could combine a regime of lawful blackmail with specific laws requiring disclosure of the kind of information the blackmail regime would otherwise inefficiently withhold, or one could combine a regime prohibiting blackmail with specific laws requiring confidentiality of the kind of information that that regime would otherwise inefficiently disclose. As between the two, the latter combination is likely to be more effective for the reason just addressed: Disclosure is inherently more verifiable than secrecy. Breaches of privacy laws are necessarily more public than breaches of disclosure laws, and the former is therefore likely to encourage more compliance than the latter. Consequently, our existing legal scheme in which blackmail is illegal is supplemented by various laws protecting privacy.(128) If these laws accurately target the kind of information that it is inefficient to disclose, then the informational outcome is likely to be more efficient than a regime of legalized blackmail.(129) A regime banning blackmail, with privacy norms and privacy laws, might achieve a dissemination of information close to the optimal level.(130)

D. Speculating on the Overall Informational Efficiency of the Blackmail Ban

Lindgren appears to concede that economic analysis demonstrates the inefficiency of commercial blackmail, where B invests in gaining information about V that, after payment, he will not disclose. Lindgren also appears to concede that economic analysis demonstrates the inefficiency of opportunistic blackmail in those cases where B will not gossip: here, permitting blackmail makes V worse off and wastes the transaction costs necessary to reach a blackmail agreement. But Lindgren forcefully criticizes the existing economic theory of blackmail for its failure to justify the ban on opportunistic blackmail in the many cases where B will gossip: here, V would prefer having the option of paying blackmail. Banning blackmail in this situation makes V as well as B worse off. Thus, Lindgren concludes that the "paradox" of blackmail continues to confound economic analysis.

This Part supplements the economic theory of blackmail to respond to Lindgren's criticism. Even assuming B desires to gossip about his discovery, within close-knit groups, the ban on opportunistic blackmail is likely to improve the distribution of information. The argument relies on the interests of TP and the greater efficiency of distributing information based on norms rather than based on the abilities of V and TP to bid against each other for secrecy or disclosure. The remaining question, however, is whether it is desirable to ban opportunistic blackmail that occurs outside of close-knit groups, as where B, V, and TP are all strangers to each other. Because there may be no norms of privacy in this context, one might imagine that the ban is inefficient.

In this final Section, however, I claim that the efficiency of the ban within close-knit groups is sufficient, in practical terms, to answer Lindgren's criticism fully. As just stated, the ban on opportunistic blackmail is clearly efficient if B will not gossip. My claim is that, outside of close-knit groups, it is probable that B will not gossip. In other words, Lindgren's criticism depends entirely on the probability of gossip, and gossip is most likely to occur within close-knit groups. Thus, the argument of this Part demonstrates the efficiency of the ban on opportunistic blackmail in precisely the situation where Lindgren's argument appears to have the greatest force.

Outside of socially connected groups, the likelihood of gossip is low. People typically do not gossip about strangers or with strangers. They enjoy gossiping about people they know with people they know.(131) The more familiar V, B, and TP are with one another, the more plausible it is that B will, absent a blackmail payment, release the critical information to TP.(132) The more probable the gossip, the more likely the group is "close-knit," because a defining characteristic of such groups is that members have a "good supply of information on past and present internal events."(133) To a significant degree then, the groups in which gossip circulates overlap with the close-knit groups that enforce norms, so that most of the opportunity for opportunistic blackmail arises among parties who are subject to common norms. Thus, Lindgren's claim that Bs gossip about their discoveries of Vs' secrets to TPs is most forceful precisely where the privacy norms argument of this Part applies.(134) Where the parties do not belong to a common group, the odds of gossip diminish substantially, and the prior argument of other commentators applies: V prefers the ban because, if blackmail is deterred, B has no reason to disclose or threaten to disclose; V will enjoy B's silence for free.(135)

Although Lindgren s criticism fails outside of socially connected groups, that does not necessarily demonstrate that the ban in this case is efficient. That Vs prefer the blackmail ban does not complete the analysis, which also depends on the consequences for TPs. Focusing on adventitious discoveries outside of close-knit groups, it is probable that TPs will not gain the information whether or not blackmail is permitted. If blackmail is banned, B will not gossip with unknown TPs. If blackmail is lawful, V will pay B not to disclose. Of course, permitting blackmail may change the costs V incurs for maintaining his secrets.(136) At this point, however, we return to the indeterminancies that plague incentive-effect analyses: It is efficient to maximize the costs of Vs secrets in some cases (as when he is concealing his undiscovered crimes or torts) but inefficient in others (as when he conceals appropriately private facts about health or sexuality).(137)

Nonetheless, economic analysis dispels the puzzling or "paradoxical" nature of the blackmail ban. The wasteful-investment argument explains the prohibition on what Lindgren terms commercial blackmail, where B discovers V's secret as the fruit of his investment. A wasteful-transaction cost argument explains the prohibition on opportunistic blackmail in situations where B would, absent the blackmail opportunity, not disclose. The existence of norms explains the ban on opportunistic blackmail within close-knit groups, where B will gossip unless paid. And gossip is likely to occur precisely within such groups. One might seek to preserve the puzzle by claiming that we could define the crime of blackmail to exclude (and thereby permit) opportunistic blackmail where the parties are not members of the same close-knit group but where B would still gossip. But even setting aside the enormous administrative costs of such a rule, there is no clear efficiency gain given the ambiguous incentive effects of such blackmail.(138) At most, one can say that there is a subset of blackmail, probably small, in which the gain from criminalization is uncertain. But this conclusion is hardly startling; given the imprecision of general statutory categories, the possibility of such imperfections is inevitable.

CONCLUSION

Rational choice analysis suggests several interesting connections between law and norms. One is the use of law to govern norms. Where prior scholarship discusses particular norms that legal rules may strengthen or weaken, this Article identifies the state's blackmail rule as a means of shaping norms generally. A legal rule banning or permitting blackmail regulates, among other things, the conditions under which members of close-knit groups can threaten to invoke certain norm sanctions, such as shame and reputational loss. The blackmail ban has ambiguous effects on the expected cost of violating norms, but more definite effects on the internalization of norms and the communication processes by which norms are refined and reformed. By preserving the public nature of norm enforcement, the ban facilitates both mechanisms. Thus, the state's decision to ban blackmail is an indirect means of universal norm governance. Ultimately, the example demonstrates that some amount of norm governance is inevitable, if only because certain legal rules will have widespread if unintended consequences for norms. If the state is to regulate norms intentionally, it should choose an optimal mix of rules either generally favoring or disfavoring norms in combination with rules facilitating or obstructing particular norms.

Rational choice analysis also identifies the conditions under which law and norms are able to influence individual behavior. Prior scholarship emphasizes the conditions in which one of these mechanisms controls behavior to the exclusion of the other, as when group norms effectively displace legal rules. Blackmail illustrates a more complex relationship, where law and norms affect behavior together and where the regulatory choice is between two different bundles of legal and norm-based rules. In particular, certain informational norms exist because of our decision to ban blackmail, but different norms would arise were we to permit blackmail. Determining the efficient legal rule requires that we compare one legal rule/norm combination to another.

In this manner, the analysis of group norms, in combination with existing economic theory, demonstrates that the entire prohibition of blackmail is efficient, and therefore explicable in economic terms. Absent norms, the blackmail ban will produce too much disclosure. Because of a market failure due to asymmetric transaction costs, a regime permitting blackmail will produce too little disclosure. Informational norms work to correct both inefficiencies, but privacy norms will command more compliance than disclosure norms. Thus, the legal rule banning blackmail, in combination with privacy norms, reaches a more efficient distribution of information than would a lawful blackmail market, in combination with disclosure norms. By insulating the information-holder from any monetary influences on his decision to withhold or disclose the secret, a blackmail ban maximizes the influence of norms on such decisions. Thus, within close-knit groups the ban is likely to be efficient, and this turns out to be of pivotal importance in defending an economic explanation of the blackmail ban. (1) Rational choice theorists in a variety of disciplines have recently taken up the phenomenon of norms, attempting to use economically inspired models to explain the origin and function of norms. See, e.g., James S. Coleman, Foundations of Social Theory, chs. 10, 11, 30 (1990); Robert Sugden, The Economics of Rights, Co-operation and Welfare (1986); Edna Ullmann-Margalit, The Emergence of Norms (1977); Robert Axelrod, An Evolutionary Approach to Norms, 80 Am. Pol. Sci. Rev. 1095 (1986); Michael Hechter, The Attainmcnt of Solidarity on Intentional Communities, 2 Rationality & Soc'y 142 (1990); Philip Pettit, Virtus Normativa: Rational Choice Perspectives, 100 Ethics 725 (1990). Jon Elster, on the other hand, uses rational choice concepts in some respects but defines norms as motivations that are not outcome-oriented. See Jon Elster, The Cement of Society 97-99 (1989). Within legal scholarship, the seminal work is Robert Ellickson, Order Without Law: How Neighbors Settle Disputes (1991). (2) See, e.g., Ellickson, supra note 1, at 123-36, 167-83,249-58 (discussing different sources of social control, the different sanctions each controller uses to enforce its rules, and the conditions determining the relative effectiveness of legal and norm based rules). (3) For recent examples outside of rational choice theory, see Donald Black, The Behavior of Law (1976); Donald T. Campbell, Legal and Primary-Group Social Controls, 5 J. Soc. Biological Structures 431, 434 (1982). (4) Indeed, where legal rules overlap with governing norms, the legal rules may be ignored. See Ellickson, supra note 1, at 40-81 (discussing the lack of influence of legal rules in disputes regarding cattle trespass and boundary fences). On the other hand, where the conditions are not appropriate for norms, legal rules still govern the relevant conduct. See id. at 82-103 (discussing the dominant influence of legal rules in disputes regarding highway collisions involving livestock). Thus, Ellickson rejects the extreme law-and-society claim that norms determine behavior to the exclusion of law ("legal peripheralism") and the extreme law-and-economics claim that law determines behavior to the exclusion of norms ("legal centralism"), and focuses on which 4controller. is more powerful in particular contexts. See id at 137-55. (5) There is no necessary tension between norm governance and the substitution hypothesis. Game theory identifies situations where desirable norms fail to arise or are underenforced; the state may then use law not merely to replace, but to strengthen norms. Came theory also identifies situations where undesirable norms arise; the state may then use law to impede norm enforcement. (6) See Robert D. Cooter, Decentralized Law for a Complex Economy: The Structural Approach to Adjudicating the New Law Merchant, 144 U. Pa. L. Rev. 1643 (1996) [hereinafter Cooter, Decentralized Law]; Robert D. Cooter, Structural Adjudication and the New Law Merchant: A Model of Decentralized Law, 14 Int'l Rev. L. & Econ. 215 (1994) [hereinafter Cooter, Structural Adjudication]. (7) See, e.g., Richard H. McAdams, Cooperation and Conflict: The Economics of Croup Status Competition and Race Discrimination, 108 Harv. L. Rev. 1003, 1064-71, 1083 (1995) (concluding that "government should obstruct. socially destructive group norms such as those underlying race discrimination); cf. Mary Anne C. Case, Disaggregating Gender from Sex and Sexual Orientation: The Effeminate Man in the Law and Feminist Jurisprudence, 105 Yale L.J. 1, 68 (1995) (advocating use of Title VII to protect effeminate men because "one of the most effective ways to improve the value of something coded feminine . . . is to make it accessible to and acceptable in men"); Cass R. Sunstein, On the Expressive Function of Law, 144 U. Pa. L. Rev. 2021, 2035 (1996) (defending the use of law to undermine dysfunctional norms promoting dangerous behaviors such as smoking and unsafe sex). (8) See, e.g., Lisa Bernstein, Merchant Law in a Merchant Court: Rethinking the Code's Search for Immanent Business Norms, 144 U. Pa. L. Rev. 1765, 1788 (1996) ("[R]ational transactors might deliberately leave aspects of their contracting relationship to be governed, in whole or in part, by extralegal commitments and sanctions."); Eric A. Posner, The Regulation of Croups: The Influence of Legal and Nonlegal Sanctions on CoUeetive Action, 63 U. CHI. L. Rev. 133, 159-60 (1996) ("[W]ealth maximizing courts should defer to groups' resolutions of internal disputes, just as courts in the administrative law context defer to agency determination."). (9) By "informational" blackmail, I mean extortion where the threat is to disclose information that is embarrassing or damaging to the victim (rather than to do something else against the victim's interests). See infra text accompanying note 18. (10) SeeJames Lindgren, Unraveling the Paradox of Blackmail, 84 Colum. L. Rev. 670, 690 (1984) (adopting the term "opportunistic blackmail" to refer to blackmail where the information the blackmailer threatens to disclose was obtained unexpectedly or accidentally, without any investment (citing Mike Hepworth, Blackmai: Publicity and Secrecy in Everyday Life 73-77 (1975))); see alsoo infra notes 84-97 and accompanying text (discussing the difficulty of justifying a ban on opportunistic blackmail). (11) See, e.g., Cooter, Decentralized Law, supra note 6, at 1684-85. Thus, norms are not merely regularities of behavior, but obligatory regularities, the deviation from which incurs disapproval and other sanctions. Philip Pettit defines a norm more carefully:

A regularity, R, in the behavior of members of a population, P, when they are agents in a recurrent situation, S, is a norm if and only if it is true that, and it is a matter of common belief that, in any instance of S among members of P,

1. nearly everyone conforms to R;

2. nearly everyone approves of nearly anyone else's conforming

and disapproves of nearly anyone else's deviating; and

3. the fact that nearly everyone approves and disapproves on this

pattern helps to ensure that nearly everyone conforms. Pettit, supra note 1, at 751. (12) Ellickson, supra note 1, at 181. See generally id at 177-82 (discussing the meaning and consequence of a group being "close-knit"). The "prospect of unavoidable future encounters" among a cluster of people gives them the ability to sanction each other; communication supplies members with information constituting the reputation of other members, which permits decisions about whether and what sanctions will be applied. Ellickson says the "vagueness" of his definition "is unavoidable. because "social cavironments are too rich to be described in terms of a few quantifiable variables." Id. at 178. He perceives the relationship between norms and close-knittedness to be continuous: "the more close-knit a group is, the better it will be able to use its informal-control system ...." Id. at 177 n.35. (13) See id. at 40-81. Ellickson also recounts how whalers in the 18th and 19th centuries used norms to resolve disputes over the ownership of whales. See id. at 191-206; see also id. at 218-19 (regarding a similar discussion of Maine lobstermen, based on James M. Acheson, the Lobster Gangs of Maine (1988)). (14) Many close-knit groups are linked only by occupation (for example, law professors, who interact at conventions, via internet, etc.), only by geography (for example, a group of long-time neighbors), or only by some interest other than occupation (for example, members of a common church or synagogue). As Ellickson notes, "a person can be a member of several close-knit groups simultaneously,. and "a group does not necessarily have to be small to be close-knit." Ellickson, supra note 1, at 182. (15) See Steven N.S. Cheung, The Fable of the Bees: An Economic Investigation, 16 J.L. & Econ. 11, 30 (1973). (16) See Elinor Ostrom, Governing the Commons: The Evoltion of Institutions for Collective Action (1990); see also Elinor Ostrom et al., Rules, Games, and Common-Pool Resources 283-300 tl994) (discussing norms associated with the role of groundwater in southern California); Laurence R. Iannaccone, Sacrifice and Stigma: Reducing Free-Riding in Cults, Communes, and Other Collectives, 100 J. Pol. Econ. 271, 289-90 (1992) (noting the role of norms and stigma in the success of religious groups). (17) Decades ago, Stewart Macaulay documented business norms in Wisconsin. See Stewart Macaulay, Non-contractual Relations in Bwiness: A Preliminary Study, 28 Am. Soc. Rev. 55 (1963). Janet Landa and Robert Cooter more recently explained the existence of "ethnically homogenous middlemen groups," ethnic minorities that tend to dominate certain industrial niches in a nation, in terms of their superior ability to use informal sanctions for contract breaches. See Janet T. Landa, A Theory of the Ethnically Homogeneous Middleman Group: An Institutional Alternative to Contract Law, 10 J. Legal Stud. 349, 355-57 (1981) (explaining that a code of ethics, embedded in kinship/ethnic relations, serves as a substitute for contract law); Robert D. Cooter & Janet T. Landa, Personal Versw Impersonal Trade: The Size of Trading Croups and Contract Law, 4 Int'l Rev. L. & Econ. 15, 21 (1984) (arguing that trading groups "perform the useful function of facilitating contracts under conditions of uncertainty"); see also Edna Bonacich & John Modell, the Economic Basis of Ethnic Solidarity 33 (1980) ("[E]thnic groups are really political-interest groups that are able to make use of moral and ritual obligations to bind their members to act in the interests of the group."). Thus, in economies with low quality legal sanctions, those groups with access to better quality informal sanctions, who can invoke the cooperative norms of their ethnic group, enjoy a comparative contractual advantage. Today, a number of contracts scholars point to informal group sanctions as a major deterrent to contractual breaches and strategic behavior, not merely in economies with underdeveloped legal systems, but everywhere. See, e.g., Lisa Bernstein, Opting Out of the Legal System: Extralegal Contractual Relations in the Diamond Industry, 21 J. Legal Stud. 115 (1992) (explaining how the contemporary New York diamond market is regulated by private norms); Posner, supra note 8, at 155 ("When members of a solidary group transact, norms and non-legal sanctions generally resolve disputes. When contingencies arise, norms allocate risks and specify means of resolution. Norms also prohibit bad faith and opportunism. The importance of maintaining a good reputation and of avoiding ostracism deters improper behavior."); see also Bryan W. Husted, Transaction Cosss, Norms, and Social Networks, 3S Bus. & Soc. 30, 53-55 (1994) (analyzing the relationship of transaction costs, norms, and social networks by comparing buyer/seller relationships in Mexico and the United States). Other research suggests that norms arise among groups of employees within a single firm or institution. See, e.g., Roderick M. Kramer, Cooperation and Organizasional Identification, in Social Psychology in Organizations: Advances in Theory and Research 244,253-59 (J. Keith Murnighan ed., 1993); Milli Laing, Gossip: Does It Play a Role in the Socialization of Nurses?, 25 Image: J. Nursing Scholarship 37, 39-41 (1993); Mike Noon & Rick Delbridge, News from Behind My Hand: Cossip in Organizations, 14 Organization Stud. 23, 31-34 (1993). (18) Model Penal Code [sections] 223.4 (1985) (Theft by Extortion); see also id. [sections] 212.5 (Criminal Coercion) ("A person is guilty of criminal coercion if, with purpose unlawfully to restrict another's freedom of action to his detriment, he threatens to: . . . (c) expose any sec.ret tending to subject any person to hatred, contempt or ridicule, or to impair his credit or business repute ....n). For a review of state statutes, see Lindgren, supra note 10, at 673-80. (19) They are what Cooter terms "cheap pain." Cooter, Deconiralized Law, supra note 6, at 1668. (20) Sociologists and anthropologists find that a love for gossip is common to many cultures. "[E]very single day, and for a large part of each day, most of us are engaged in gossiping. I imagine that if we were to keep a record of how we use our waking-time, gossiping would come only after 'work'--for some of us--in the score.. Max Cluckman, Cossip and Scandal, 4 Current Anthropoly 307, 308 (1963). For further discussion, see Jorg R. Bergmann, Discreet Indiscretions: The Social Organization of Gossip (1993); Good Gossip (Robert F. Coodman & Aaron Ben-Ze'ev eds., 1994); Ralph L. Rosnow & Gary A. Fine, Rumor and Gossip (1976); PaTRICIA M. SPaCKS, COSSIP (1985); Sally E. Merry, Reshinking Cossip and Scandal, in 1 Toward a General Theory of Social Control 271 (Donald Black ed., 1984). (21) See Noon & Delbridge, supra note 17, at 30-31 (discussing factors that affect whether and how particular gossip is transmitted). After all, what we mean by a "close-knit" group is one in which "the information pertinent to informal control circulates easily among [members]." Ellickson, supra note I, at 177-78. Circulation is not, however, strictly necessary for all norms. Some behavior-the clothing one wears, for example-may be so public that such norms are easily enforced even if no one discloses to others when they observe a violation. But it is rare that everyone in the group directly observes a violation; thus, norm enforcement is usually enhanced if observers also disclose the violation to others who can also sanction the violator. Where discovery is rare, gossip may be necessary to the existence of the norm. See Cluckman, supra note 20, at 312-15; A.L. Epstein, Cossip, Norms, and Social Norworis, in Social Networks in Urban Situations 117 ("Clyde Mitchell ed., 1969). (22) Disclosure is not guaranteed. Even without blackmail, the discoverer of the violation will in some cases have selfish reasons for concealing the information. For example, he may fear retaliation by the norm violator. But disclosure is the common result. (23) In theory, the state could decriminalize blackmail but not enforce blackmail contracts. In evaluating the effect of legalization, however, I assume that blackmail contracts would be enforceable. Even if they were not, decriminalization would itrelf raise the expected returns of blackmail, increasing its quantity and producing the effects discussed in this Part. (24) See Richard A. Epstein, Blacimail, Inc., 50 U. Chi. L. Rev. 553, 562 (1983) (noting that after legalization, "there would . . . be an open and public market for a new set of social institutions to exploit the gains from" blackmail). Civen the existence of norms, this claim is contestable. There are norms against blackmail, and it is conceivable that the norms are so powerful and stable that the removal of criminal sanctions would not affect the quantity of blackmail. But it is highly probable that blackmail would substantially increase. The private gain from other property crimes, such as theft, is often sufficiently large that norm sanctions are inadequate deterrents, especially for those who take property from victims outride their social groups. Unlike conventional theft, where the victim is likely to announce the violation of an antitheft norm, the blackmail victim is likely to conceal his blackmailer's norm transgression, so as to preserve the secret for which he is being blackmailed. Thus, norms are less able to deter blackmail than simple theft. Absent substantial criminal penalties, the expected return from blackmail is likely to be positive in many cases.

Moreover, the reported cases of blackmail prosecutions demonstrate that norms are not always sufficient to deter blackmail. See Richard A. Posner, Blachmail, Privac,, and Freedom of Contract, 141 U. Pa. L. Rev. 181?, 1841-47 (1993) (finding 124 published opinions involving a "blackmail" prosecution in this century). Although such prosecutions are relatively rare, their scarcity does not imply that criminal sanctions are unimportant. As Posner explains, id. at 1857-40, the blackmail prohibition works primarily by arming victims with a potent counterthreat: "Expose my secret and I'll expose your blackmail." Thus, Posner concludes that reported cases are relatively rare because, given criminal sanctions, successful blackmail is relatively rare. See id. at 1841. I further discuss the weaknesses of antiblackmail norms in infra note 126. (25) One might respond that legalizing blackmail would not decrease disclosure, but would merely increase the price paid for disclosure. Those who value the information could pay for it, and they would pay if they valued disclosure more than the victim valued secrecy. This solution, however, usually fails. As I explain in considerable detail, a blackmailer faces far lower transaction costs in dealing with the traditional victim-who already knows the secret--than with those who would value the information if they knew of its existence--that is, the other group members. See infra notes 101-09 and accompanying text. (26) This statement is not always correct. If an individual has "internalized" the norm, see infra notes 36-42 and accompanying text, then he will suffer some cost--guilt--even if his violation is not detected. But I defer discussing how blackmail affects internalization until Part I.B and focus here on how blackmail affects external sanctions. Even when a group member has internalized a norm, he will violate it if the benefits exceed the internal costs plus the expected external sanction. (27) Many scholars have made this point in the context of crime. The most elaborate analysis appears in Jennifer C. Brown, Blackmail As Private Justice, 141 U. Pa. L. Rev. 1935, 194-49 (1993). (28) If the norm violation is also a crime, V will commit the violation only if b > P(sc + cp), where cp is the criminal punishment. The arguments I make for why bp < se also explain why bp < sc + cp. Cf. Steven Shavell, An Economic Analysis of Threats and Their illegality: Blachmail, [Extortion, and Robbery, 141 U. Pa. L. Rev. 1877, 1882 (1993) (arguing that when a victim seeks to conceal criminal activity, the blackmail price will tend to be lower than the punishment the victim will otherwise face). (29) The possibility that others besides B may discover V's secret and either disclose it or demand a blackmail price does not mean that V will pay more, on average, than se. V will tend to discount the value of B's promise not to disclose by the probability that others besides B might disclose. Consequently, the possibility of independent discovery should not systematically affect the relative values of se and bp. For example, if there is a 10% chance of independent discovery and disclosure, V will not pay B more than O.9se. Suppose V pays B 0.9sc and that the independent discoverer discloses V's norm violation. In one out of 10 cases, V bears a cost of 1.9sc, while in nine out of 10 cases, V bears a costs of 0.9se. The average is still 1.0sc. (30) B's reservation price is often just the value he places on disseminating the information through gossip. There is not likely to be any pecuniary return to B for his gossip. See infra notes 101-09 and accompanying text. (31) Note the asymmetry: When V is the superior bargainer, he will pay less than se, but when B is the superior bargainer, he will not extract more than sc. When B overestimates V'5 reservation price and demands a payment higher than sc, V will refuse to pay; but when B underestimates V's reservation price and demands a payment lower than sc, V will accept the deal. (32) There is possibly a third reason. V will pay B less than sc if he believes B might avoid full liability for breaching his promise not to disclose V's secret. For example, if V believes he cannot adequately prove his full damages, nor be certain a court will enforce a liquidated damages provision, then he may fear that B will subsequently threaten to breach the contract unless paid further sums. Even if V can prove damages, he may fear that B lacks the wealth to pay them. If V i5 unsure whether B would pay all of V's damages (equal to sc) upon breach, V will accordingly discount the value of B's promise.

These administrative costs will be reduced, possibly to a negligible level, if the legalization of blackmail creates a large blackmail industry, what Richard Epstein has labeled "Blackmail, Inc." Epstein, supra note 24. In such an industry, firms would compete to establish reputations for keeping their promises and threats--for concealing secrets (without further demands) when paid the blackmail price and disclosing secrets otherwise. With perfect information about these reputations, V would have no reason to expect that Blackmail, Inc. would breach its promise or be unable to pay damages if it did. When someone not in the business discovered information suitable for blackmail, he would maximize his return by selling the information to Blackmail, Inc., which would actually make the contract with V.

Reputational information is, however, not perfect. Moreover, where an "amateur" sells his secret discovery to Blackmail, Inc., the firm will face the same difficulties discussed above in preventing the amateur from disclosing the secret or selling it again: Blackmail, Inc. may not be able to prove its full damages or recover the damages it can prove. Blackmail, Inc. would often have reasons not to disclose its informational source to V, so even when one of its professional agents obtained the information, V will worry that the source is a less controllable amateur. Thus, these administrative problems may provide another reason that the amount V pays B (even Blackmail, Inc.) will fall below sc. (33) The indeterminacy is more severe because of additional effects not discussed in the text. For example, by decreasing gossip, blackmail may in one way decrease the probability of detection. I have in mind situations where discovering a norm violation requires interpreting ambiguous information. Suppose 0 observes group member V behave furtively or erratically. Absent the opportunity for blackmail, O will share this information with other group members through gossip. By combining data, the group may jointly discern norm violations no individual would have discovered. The opportunity to blackmail, however, creates an incentive not to share information. Each individual will hold back his suspicions in the hope of independently discovering evidence of a norm violation that will allow him alone to blackmail V. Croup discovery forces one to share blackmail proceeds with others or, if the discovery is widespread, eliminates blackmail leverage entirely. Legalizing blackmail thus creates a collective action problem: the group benefits if individuals share their knowledge, but each individual now benefits from hoarding information.

There is also a way in which allowing blackmail would increase the sanction for violating norms. When the discoverer of a norm violation is not a member of the norm violator's group, he may detect the violation, but he is not likely to gossip. As a general rule, people do not gossip about strangers or with strangers. People gossip with people they know about other people they know, usually when the gossip recipient also knows the subject of the gossip. See Bergmann, supra note 20, at 67 (U[T]he subject of gossip must be an acquaintance of the gossip recipient or at least be known to him indirectly through intermediaries or local 'fame,' because the news he hears is personally relevant for him only if it is not about a complete stranger."); Sally Yerkovich, Gossiping As a Way of Speaking, J. Comm., Winter 1977, at 192, 196 ("Information, no matter how salient or scandalous, isn't gossip unless the participants know enough about the people involved to experience the thrill of revelation."). With extragroup discovery, sc is effectively zero. Thus, in this situation, the blackmail transaction replaces zero punishment with a blackmail price. Permitting blackmail gives strangers an incentive to use negative information they discover--that is, to blackmail--in situations where they have no incentive to gossip. (34) Suppose, for example, that without blackmail, P is already very high (near 1.0). It might still be the case that the norm is underenforced if b is sometimes higher than sc. In this case, there are two reasons to think that permitting blackmail will have very little positive effect on P. First, there is little reason to invest in discovering violations that are already nearly certain to be discovered; the expected blackmail return is low. Second, when P is quite high, it can only increase so much. In this circumstance, the decrease in sc may dominate. We might predict that permitting blackmail would reduce the expected cost of violating those norms in which the probability of detection is already high. Cf. supra note 32 and accompanying text (arguing that the decrease in sc is greatest for the group's most important norms). (35) With risk aversion, one could assert that an increase in P and an equivalent decrease in sc would decrease norm compliance because risk is reduced. But there is still no way of knowing that the increase in P is "equivalent" to the decrease in sc. (36) See Robert D. Cooter, Law and Unfied Social Theory, 22 J.L. & Soc'y 50, 63 (1995) ("[A] customary norm emerges in a community when it is internalized by enough members.") [hereinafter Cooter, Social Theory]; see also Emile Durkheim, The Division of Labor in Society 68-87 (W.D. Halls trans., 1984); Max Weber, The Protestant Ethic and The Spirit of Capitalism 27 (Talcott Parsons trans., 1958); Robert C. Ellickson, Bringing Culture and Human Frailty to Rational Actors: A Critique of Classical Law and Economics, 65 Chi.-Kent L. Rev. 2S, 46 (1989). (37) See Cooter, Decentralized Law, supra note 6, at 1665 ("[S]omeone who has internalized a norm feels guilt from violating it and pride from obeying it."). (38) See id. at 1666. (39) Because norms are internalized only where "private incentives for signaling align with a local public good," Cooter believes that internalized norms tend to be efficient. See Cooter, Structural Adjudication, supra note 6, at 224. (40) For a review of conformity experiments, see Serge Moscovici, Social Influence and Conformity, in 2 The Handbook of Social Psychology 347, 350 (Gardner Lindzey & Elliot Aronson eds., 3d ed. 1985) ("Hundreds of experiments have demonstrated that people can be made to state and to think the opposite of what they see and believe to be the truth."). As Cooter's model requires, the fact that such judgments are unanimous is particularly important to measuring a conformity effect. (41) See id. at 363, 368-70. (42) See id. at 375.

[T]he individual continues to conform somewhat even when he is alone and

anonymous. All in all, there is never mere compliance or mere change in

attitude. And as we know because of the introduction of the cognitive

dissonance theory, a forced compliance may result in an authentic change

in attitude. This change results because individuals try to reduce the

discrepancy between their actions and their thoughts, their public end their

private opinions. Pascal recognized the fact long ago: Pray, and faith will

follow. Id. at 393. (42) See Cooter, Decentralized Law, supra note 6, at 1669. (43) Sociologist John Braithwaite observes:

Through listening to and participating in secretive gossip directed at others

we learn the circumstances by which people suffer loss of reputation

through gossip. . . . Children need to learn about the evil of murder, rape,

car theft, and environmental pollution offences through condemnation of

the local butcher or the far away image on the television screen. But the

shaming of the local offender known personally to children in the

neighborhood is especially important, because the wrongdoing and the

shaming are so vivid as to leave a lasting impression. . . . The evil of acts

beyond the immediate experience of children is more effectively communicated

by shaming than by pure reasoning. John Braithwaite, Crime, Shame, and Reintegration 76-78 (1989); see also Gary A. Fine, Social Components of Children's Gossip, J. Comm., Winter 1977, at 181, 182 ("Numerous studies of children have emphasized the large amount of social conformity that occurs throughout the growing up period. . . . One of the determinants of this conformity is through the normative regulation of gossip." (citation omitted)). See generally Jerome Kagan, The Nature of the Child 14-49 (1984) (discussing how violations of standards result in emotional experiences that may create internal standards within the child). (44) As one commentators notes:

A major component of professional socialization is internalization of

the norms and values of the professional and work culture. . . .

Friedson . . . submits that the use of hearsay, stories and gossip among

professionals is a strong form of social influence. Traditional morals and

ideals can be made public through gossip and stories. . . . This type of

communal scrutiny and sanctioning also occurs in nursing. During change

of shift report, experienced nurses verbalize their concerns regarding

standards of care, educating neophytes to what is, or is not, acceptable

nursing practice . . . . Laing, supra note 17, at 40-41 (citing Eliot Friedson, Doctoring Together: A Study of Professional Social Control 142-45 (1980)). (45) See Thomas Gabor, Everybody Does It! Crime by the Public 183-84 (1994) (noting that a frequent justification for employee theft is that such conduct is commonplace). (46) The distinction between norms and meta-norms is a useful one. Internalization probably occurs at a general level--one is likely to internalize the general obligation to be a "good friend" or to treat others "fairly" long before internalizing the specific behaviors that, at a given time and place, are thought to constitute good friendship or fair treatment. This distinction also reconciles the apparent timidity of norms with the stability of internalization. The obligation to "do one's share," to be a "good friend," or to be treat does not change much over time, but the "social meaning. of each of these obligations does. Cf. Lawrence Lessig, The Regulation of Social Meaning, 62 U. Chi. L. Rev. 943 (1995) (discussing the influences of "social meaning" on individual behavior). (47) Of course, some norm violations will be discovered simultaneously by such a large number of people that no opportunity for blackmail will arise. Thus, blackmail will not eliminate public enforcement. Nonetheless, norm violators will seek to conceal their violations, so that many discoveries will represent blackmail opportunities, meaning lost opportunities for public enforcement. (48) Cf. Brown, supra note 27, at 1970-71 (noting that, regarding criminal offences, the blackmailer cannot perform the "important task" of "declaring societal norms and labeling as 'criminal' behavior that runs afoul of them"). (49) I am assuming that the norm is one in which the existence of violations is typically public, although the identities of the violators is not. For example, people know the extent to which others are violating an antilittering norm or a water-conservation norm, although they do not typically know the identities of the violators. (50) See Braithwaite, supra note 43, at 75. He further notes:

In summary then, shame operates at two levels to effect social control.

First, it deters criminal behavior because social approval of significant others

is something we do not like to lose. Second, and more importantly, both

shaming and repentance build consciences which internally deter criminal

behavior even in the absence of any external shaming associated with an

offence. Shaming brings into existence two very different kinds of

punishers--social disapproval and pangs of conscience. Id. (51) See id. at 81. (52) When confronted by the group, norm violators may realize that what seemed sound reasoning in private will fail to persuade anyone else. To hasten reintegration, the violator eventually considers repenting. But the violator will more easily convince others he is repentant if he really is repentant. Those who achieve genuine remorse have begun to internalize the norm. (53) It is possible, of course, that the blackmail victim will view his plight as something he deserves, and blackmail will cause him to feel remorse and internalize the norm. This result is more likely, however, when the punishment is overt and communal. (54) First, any efficiency claim is merely that the norm benefits the group in which it arises; thus, it is entirely possible that the norm benefits the group by imposing a larger cost on another group. See Ellickson, supra note I, at 169 ("[N]orms that add to the welfare of the members of a certain group commonly impoverish, to a greater extent, outsiders to that group."); Cooter, Structural Adjudication, supra note 6, at 225 ("Communities often develop norms that benefit their members at the expense of members of other communities."); McAdams, supra note 7, at 1064-71 (noting that norms of racial discrimination generate status welfare for the majority group at the expense of subordinate minority groups). Second, even for the group in which it arises, if the process of norm change is evolutionary, the norm may only reach a "local maximum." See Cooter, Structural Adjudication, supra note 6, at 225. Evolution occurs incrementally. But if the first few steps toward a global maximum cause a decline in adaptive fitness (group welfare), then the organisms (here, groups) making such a change may die out (be abandoned) before they reach the point of increasing returns. Of course, human groups are not limited to evolutionary change; they may change their norms consciously, "jumping" from a local to global maximum. But a nonevolutionary mechanism for producing norms requires information. Individuals must accurately perceive a harm, its magnitude, and its causes. Otherwise, group members may enforce a norm that causes rather than corrects the problem, or that imposes higher costs than the harm the norm reduces. For a more detailed discussion of the barriers to norm efficiency, sec Eric A. Posner, Law, Economics, and Inefficient Norms, 144 U. Pa. L. Rev. 1697 (1996). (55) See Barbara Yngvesson, The Reasonable Man and the Unreasonable Gossip: On the Flexibility of (Legal) Concepts and the Elasticity' of (Legal) Time, in Cross-Examinations: Essays in Memory of Max Gluckman 133, 134 (P.H. Gulliver ed., 1978) ("Gossip . . . serve[s] to clarify politically problematic issues and to provide a publicly acceptable, 'straightforward' explanation of how a breach is to be handled."). Yngvesson compares the use of courts and gossip to adjudicate disputes and argues that they differ less than is usually supposed:

All of these procedures--presentation of a case by litigants in court,

disposition of the case by a judge in court, plea-bargaining over the charge

prior to arraignment or trial, evaluation of a case by a drug screening board,

and gossip in the community--are characterized by a degree of rationality,

to the extent that they are goal-directed and involve fitting unique events

into a more general framework in terms of which these events can be

viewed as 'reasonable or 'unreasonable.' But the decisions reached surely

are shaped by irrational elements as well. whether they are reached in court . . . or

over a coffee table. Id. at 153-54; see also Noon & Delbridge, supra note 17, at 28 ("A person may gossip to seek to change group values, thus gossip can assist the individual to impact upon the group, and perhaps change its order or structure."). For further discussion of how individuals attempt to change norms, see Lessig, supra note 46; Allan Gibbard, Norms, Discussion, and Ritual. Evolutionary' Puzzles, 100 Ethics 787 (1990); Edna Ullmann-Margalit, Revision of Norms, 100 Ethics 756 (1990). (56) For example, a religious group's norm bars sex outside of marriage, and the facts show that two unmarried members had sex. (57) Suppose, for example, a norm requiring one to contribute a "fair share" to public goods, especially in time of emergency-the kind of norm that may explain why Mississippi River Valley residents constructed sandbag walls during the floods of 1993, despite the apparent incentives to free-ride. Two complications impede the application of this norm: There are a variety of different tasks that need to be done and people have a variety of abilities to do them. One imagines the norm is to contribute something like "a fair share" given one's abilities. In defining their "fair share," residents must evaluate what others contribute while making accommodations for each other's physical frailty. Application of the norm is problematic. In addition, deciding where to pile sandbags may require resolution of conflicting norms, if different norms favor efforts to save different structures and all norms cannot be accommodated. The circle of gossip is the "court" in which norm violations are "tried." Croup members use gossip to discover how much work other members did and also to work out how much work was "enough" under the circumstances. Members also determine whether decisionmakers correctly resolved conflicts between competing norms. (58) Yngvesson elaborates:

[People use gossip] in situations where rules are ambiguous . . .. [P]eople

create moral and political rules that are inherently ambiguous and flexible,

and . . . this characteristic of rules is used in situations where 'real world'

events and actions must be explained and justified. In courtroom hearings,

the fitting process is done in public . . .. Gossip does the same thing in

private, and thus is appropriate in cases where the 'fit' between rules and

behaviour may be particularly problematic.

Yngvesson, supra note 55, at 153. (59) Cf. Brown, supra note 27, at 1971-72 (noting that blackmail deprives society of criminal prosecutions that help to "clarify" law and identify "laws that may need modification"). (60) See Owen M. Fiss, Against Settlement, 93 Yale L.J. 1073, 1085 (1984) (arguing that, by deciding cases, judges "explicate and give force to the values embodied in authoritative texts such as the Constitution"). (61) Sec Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States 17 (1970). (62) For example, after a quality decline, consumers will use the "voice" option if they think that the probability that the producer will return to the preexisting quality level, times the value they place on that preexisting quality over the next best alternative, is greater than the costs of exit. Sre ;d. at 37-39. (63) See id. at 4. (64) See id. at 30. In conformity experiments the power of the group to make others "go along" dramatically declines if there is a single dissenter from an otherwise unanimous consensus. See Moscovici, supra note 40, at 363 ("[I]f even one confederate has been instructed to disrupt the unanimity by giving a different response, the number of conformist responses drops precipitously . . .. [U]nanimity . . . is more important than the mere number of persons adopting a common response."). For a discussion of conformity experiments, see supra note 40 and accompanying text. (65) There are other ways to characterize the cost of criticizing a norm. For example, the critic may be viewed as defending and condoning norm violators, thereby failing to contribute sufficiently to sanctioning the violation. In other words, there may be an independent norm against criticism of a certain norm. (66) Let [delta] [P.sub.r] be the change in the probability of reform caused by an individual's criticism, rb be the benefit the critic would receive from reform, [delta] [P.sub.s] be the change in the probability of the critic receiving the sanction, and sc be the sanction cost. The individual will criticize when [delta] [P.sub.r] (rb) > [delta] [P.sub.s] (sc). (For an altruist, rb will include some or all of the benefit other group members would gain from norm reform. Thus, altruism increases an individual's willingness to criticize a norm he believes is inefficient.) (67) Lawrence Lessig has made the same point in greater detail. See Lessig, supra note 46, at 997-99, 1006-07.

Social meanings act to induce actions in accordance with social norms, and

thereby impose costs on efforts to transform social norms. They present,

then, a particularly harsh collective action problem, for not only is there

little incentive for an individual to contribute to a new collective good,

but there is a punishment--the cost of deviance--for any individual who

wishes to contribute to a new collective good; that is, to a new social

meaning.

Id. at 998. Although groups use norms to induce contributions to public goods, norms may be ill-suited to ensuring sufficient criticism of norms. There may be a trade-off between the intensity of norm enforcement and a willingness to listen to arguments for abandoning norms. Some groups may manage to cultivate tolerance of norm criticism, but many groups enforce norms with sufficient zeal that criticism will be risky for the critic and therefore undersupplied.

In theory, norm criticism could be a public "bad" rather than a public good, if members criticize and "reform" an efficient norm. But even with imperfect information, I assume sufficient rationality that criticism is much more likely to change inefficient norms than efficient norms. It is difficult to persuade people that a norm is contrary to their interests even when it is; people tend to interpret ambiguous information as supporting their existing beliefs. See Thomas Gilovich, How We Know What Isn't So: The Fallibility of Human Reason in Everyday Life 49-72 (1991). That a norm actually benefits the group will make it even more difficult to persuade people otherwise. (68) Because blackmail raises the probability that a norm violation will be discovered but lowers the cost the violator incurs upon discovery, the net effect is indeterminate. (69) Absent blackmail, the individual's expected benefit from reforming a norm is [delta] [P.sub.r](rb). See supra note 66. Let k be a multiplier representing the effect of blackmail on that benefit. The term k is less than 1.0 if the net effect of blackmail is to lower the expected cost of violating norms and greater than 1.0 if the net effect is to raise the expected cost. In other words, a member's benefit from reforming an inefficient norm increases if blackmail makes violating the norm more costly, but decreases if blackmail makes violating the norm less costly. Absent blackmail, the individual's expected cost from reforming a norm is [delta] [P.sub.s](sc). Let [[pi].sub.bm] represent the value of all future profits an individual expects to receive from blackmailing others who violate a norm. Thus, the new cost of criticizing the norm is the value of these profits multiplied by the probability that the individual's criticism will reform the norm and prevent him from blackmailing violators. So, with blackmail, an individual will now criticize the norm if [delta] [P.sub.r](rb)k > [delta] [P.sub.s](sc) + [delta] [P.sub.r]([[pi].sub.bm] ). The textual point is that the effect of blackmail on the benefits of criticism is uncertain, because the value of k may be greater or less than 1.0, but the effect of blackmail on the costs of criticism is definitely positive, because [[pi].sub.r] > 0. Note that norm criticism will decline even if only some but not all individuals expect to reap blackmail profits (as would be the case if some but not all individuals have so internalized a norm against blackmail that they will forgo any blackmail opportunity). (70) The textual assumption represents the only case in which blackmail might possibly increase norm criticism. Consider the other two cases. First, assume blackmail has no net effect on the expected costs of violating norms (that k = 1). See supra note 69. In this case, the benefits of criticizing a norm are exactly the same with or without blackmail, but because criticism risks losing future blackmail profits, blackmail unambiguously decreases criticism of inefficient norms. Second, assume blackmail decreases the expected costs of violating norms (that k < 1 ). Here, blackmail decreases the enforcement of both efficient and inefficient norms. Because the benefits of norm criticism decline while the costs increase, blackmail again unambiguously decreases criticism of inefficient norms. The text discusses the third case, where blackmail increases the expected costs of norm violations (k > 1). (71) If an individual stands to gain nothing from norm repeal, he might be the most persuasive critic, but he also might have no reason to bother criticizing the norm. So the most persuasive critics are likely to be those who gain the least return from norm repeal, but who still gain some positive return. (72) A caveat is necessary. If an individual has sufficiently internalized a norm against blackmail, he will not blackmail others and will expect to receive no blackmail profits. If legalizing blackmail then increases the benefits of reforming inefficient norms, legalization will make such individuals more likely to criticize the norm. And because such group members include those who personally gain little from repeal, some of their criticism will be persuasive. Thus, notwithstanding the textual argument, the net effect of legalization on criticism is indeterminate if two conditions hold: legalizing blackmail increases the net cost of violating norms and a substantial pool of group members will, after legalization, be unaffected by the prospect of blackmail profits. Obviously, both conditions are empirically possible.

I stated my reasons for doubting the first proposition in Part I.A. As for the second, I think that most people would be able to rationalize to themselves their lawful opportunistic blackmail just as a great many people today rationalize their unlawful shoplifting, tax evasion, insurance fraud, employee pilfering, "hotel linen lifting,. and other "acceptable" forms of theft. See Gabor, supra note 45, at 73-97 (discussing how amateur thieves seek to justify their crimes). Also, the existing norm against blackmail is partly supported by its criminal prohibition, and is likely to be weakened when this symbol of condemnation is repealed. (73) This would be true if the sanction were already the highest the group inflicted (for example, ostracism) or if the sanction is simply not calibrated to distinguish between repentant and unrepentant violators. Even if lack of repentance is punished more severely, the violator may be able to first criticize the norm and then repent. Unless the norm is calibrated to distinguish between immediate repenters and those who repent after initial defiance, the criticism still costs the victim nothing, and may generate a critical dialogue. Even if the norm is so calibrated, the violator might be able to avoid the extra punishment if he constrains his criticism to the norm's application to situations like his, while continuing to endorse the norm generally. (74) Because blackmail purchases secrecy, it places V in the position of those whose violations remain completely undiscovered: norm criticism arouses suspicion that they are violating the norm and raises the risk of (additional) discovery. As long as the norm violator's reputation remains unsullied (he is not already bearing se), his marginal cost for criticizing the norm is positive. (75) Even if there were altruists willing to criticize the norm, see supra note 66, the public enforcement of the norm may provide the occasion and the forum at which an altruist would most effectively communicate his criticism. If blackmail decreases or eliminates public enforcement, it may also decrease or eliminate the opportunity for effective altruistic criticism. (76) See Posner, supra note 8, at 136 ("[W]hen groups are sufficiently cohesive, or solidary, and pursue goals that are consistent with the state's, transferring resources to them is a more efficient method for obtaining those goals than conventional regulation of individual action."). (77) See id. at 148-50 (analyzing the effects of different types of category-based rules on the cooperation-defection differential). (78) See Ellickson, supra note I, at 167 (hypothesizing that "members of a close-knit group develop and maintain norms whose content serves to maximize the aggregate welfare that members obtain in their workaday affairs with one another"). For similar theses, see Coleman, supra note 1, at 249-58; Cooter, Structural Adjudication, supra note 6. Regarding their exceptions to this claim, see supra note 54. (79) For pessimistic views, see Russell Hardin, One For All: The Logic of Group Conflict 72-106 (1995); Posner, supra note 54. (80) The analysis depends on the reasons for the inefficiency of norms. If norms are inefficient from the perspective of the group (and therefore socially inefficient as well), then the analysis of Part I.C argues for the blackmail ban. Because blackmail impedes one of the group's mechanisms for reforming inefficient norms (voice), the greater the tendency of norms to be inefficient (from the group's perspective), the more important it is to preserve that reform mechanism. On the other hand, if norms are inefficient because they typically benefit the group in which they arise at the greater expense of society, then the analysis of part I.C argues against the blackmail ban because voice will make groups more effective at rent seeking. (81) Jim Lindgren first referred to the problem as the "paradox" of blackmail. See Lindgren. supra note 10, at 671. Lindgren was not the first to notice the difficulty of distinguishing blackmail threats from lawful threats. See, e.g., Douglas H. Ginsburg & Paul Shechtman, Blackmail: An Economic Analysis of the Law, 141 U. Pa. L. Rev. 1849 (1993) (first circulated in unpublished form in 1979); Eric Mack, In Defense of Blackmail, 41 Phil. Stud. 273, 273-74 (1982); Jeffrie G. Murphy, Blackmail: A Preliminary Inquiry, 63 Rionist 156, 156-57 (1980). But Lindgren's formulation of the problem spurred many other theorists to attempt an explanation of the blackmail ban. Sec, e.g., 4 Joel Feinberg, The Moral Limits of the Criminal Law: Harmless Wrongdoing 238-76 (1988); Ronald H. Coase The 1987 McCorkle Lecture: Blackmail, 74 Va. L. Rev. 655, 656 (1988); Symposium, Blackmail, 141 U. PA. L. REv. 1565 (1993). (82) Thus, my goal here is solely to demonstrate how a norms analysis usefully supplements an existing economic theory of blackmail. If I am right, my analysis answers the most telling criticism of the economic approach to blackmail. Beyond this, however, I do not attempt to evaluate the relative merits of the economic and noneconomic theories of blackmail. (83) See Lindgren, supra note 10, at 690 (citing Hepworth, supra note 10, at 75-76). (84) See, e.g., Coase, supra note 81, at 671; Ginsburg & Shechtman, supra note 81, at 1859-65; Posner, supra note 24, at 1820. (85) Blackmail is like theft in this regard. See Posner, supra note 24, at 1820. (86) See, e.g., Brown, supra note 27, at 1943-49 (arguing that legalizing blackmail of criminals would increase deterrence of crime); Posner, supra note 24, at 1821-27 (finding ambiguous effects); Shavell, supra note 28, at 1892 (arguing that banning blackmail of criminals has ambiguous effects on deterrence, because "the likelihood of punishment should fall even though the magnitude should rise"). (87) See Shavell, supra note 28, at 1891-92. (88) For example, Shavell compares the effect of permitting blackmail to a system of rewards for those who provide evidence to the state and also considers the special problem of exempting law enforcement officers from a regime allowing blackmail of criminals. See Shavell, supra note 28, at 1899-1901. Brown discusses the effect of permitting blackmail with and without rules imposing a duty to report criminal activities, and also in comparison to Shavell's system of bounties. See Brown, supra note 27, at 1943-49. Among other things, Posner distinguishes between blackmail of criminal acts already punished by the state and criminal acts not yet punished. See Posner, supra note 24, at 1821-27. (89) See Posner, supra note 24, at 1827 ("As a costly and apparently sterile redistributive activity, blackmail fits the economic definition of a common law crime. The speculative argument that blackmail might serve a socially productive role . . . does not justify removing it from the prohibited category, when the opposite argument . . . is equally plausible." (footnote omitted)). (90) See Lindgren, supra note 10, at 690-91; James Lindgren, Blackmail: An Afterword, 141 U. Pa. L. Rev. 1975, 1984-86 (1993) [hereinafter Lindgren, Blackmail: An Afterword]; James Lindgren, Blackmail: On Waste, Morals and Ronald Coase, 36 UCLA L. Rev. 597, 600-03 (1989). Lindgren uses the argument to criticize economic theories of the blackmail ban and strengthen the case for his noneconomic theory.

In making his point, Lindgren borrows Hepworth's four-part taxonomy. See Hepworth, supra note 10, at 73-77. The two other categories of blackmail are: "entrepreneurial" and "participant." Both mean the blackmailer acquired the blackmail information by participating with the victim in the behavior that the victim wishes to conceal. "Entrepreneurial" blackmail means the participation was done with the purpose of setting up a blackmail opportunity; "participant" means the opportunity for blackmail did not induce the blackmailer's participation. For simplicity, I use the term opportunistic to include participant blackmail and commercial to include entrepreneurial blackmail. (91) See generally Anthony T. Kronman, Mistake, Disclosure, Information, and the Law of Contracts, L.J. Legal Stud. 1 (1978) (explaining why contract law only sometimes imposes a duty to disclose on a party with greater information). (92) In a similar vein, consider Richard Posner's explanation for the fact that blackmail is defined to require a thrmt. A person does not commit blackmail unless he threatens to disclose another's secret; receiving money for keeping silent is not blackmail if one did not first threaten to disclose. Posner defends this distinction by claiming that, if one cannot threaten to disclose information damaging to others, one is not likely to invest in acquiring such information. See Posner, supra note 24, at 1836. Thus, the rule is narrowly tailored, according to Posner, only to prevent sterile investment. Lindgren's point is that the economic explanation of blackmail implies that we should tailor the rule slightly further to permit threats to disclose information if there was no investment to acquire the information. (93) A possible response is the difficulty of distinguishing between commercial and opportunistic blackmail. Given the opportunity for concealing one's investment, the administrative and/or error costs of making such a distinction may be high. But I find this explanation unsatisfying. First, if Kronman is right, the common law makes just such a distinction--between commercial and opportunistic information discoveries--in contract law. See Kronman, supra note 91, at 18. Yet criminal law is usually more willing to bear administrative costs to accord individualized justice than civil law. And from an economic perspective, it is generally more important to narrowly tailor punitive sanctions than remedial ones. Second, if Richard Posner is right, blackmail law requires a threat because, ex ante, sterile investment will occur only if blackmailers can make threats. See Posner, supra note 24, at 1836. But the administrative costs of distinguishing "implied" threats from nonthreats are probably as great as those of distinguishing commercial from opportunistic blackmail. More generally, though administrative difficulties are undoubtedly powerful explanations in some contexts, there is a danger of overusing the explanation whenever the substantive analysis breaks down. Among other things, this Article offers a way of explaining the ban on adventitious blackmail that is not highly dependent on administrative costs. (94) See, e.g., Ginsburg & Shechtman, supra note 81, at 1875-76 (arguing that an adventitious discoverer has no reason to incur the expense of publication unless the is looking to future opportunities for blackmail"). If the costs of disclosure cause adventitious discoverers to keep silent, then the subjects of the information do not require a blackmail contract to secure their secret. (95) See Lindgren, Blackmail: An Afterword, supra note 90, at 1984-86 (noting that people frequently gossip without receiving any tangible benefits in return). (96) See supra note 20. (97) Shavell responds to Lindgren's challenge by asserting the incentive effects on the victim: "[T]o avoid being blackmailed by workmen or others who might by chance be present, potential victims will exercise excessive precautions or reduce their level of innocent, yet embarrassing, activities." Shavell, supra note 28, at 1903. But Lindgren correctly claims that this kind of analysis assumes that, absent the opportunity for blackmail, the embarrassing information will not be disclosed. See Lindgren, Blackmail: An Afterword, supra note 90, at 1984. If people enjoy gossip, however, potential victims will invest in avoiding such embarrassments even if opportunistic blackmail is banned. (98) There is a substantial body of economic criticism of privacy claims, based on the insight that privacy facilitates fraud. Sec, e.g., Richard A. Posner, The Economics of Justice 232 (1981) ("[S]ocial, like business, dealings present opportunities for exploitation through misrepresentation."); see also Symposium, The Law and Economics of Privacy, 9 J. Legal Stud. 621 (1981). But there clearly remains a category of private facts the disclosure of which is inefficient. For example, Richard Posner distinguishes between "discrediting" and "embarrassing" publicity, the latter consisting of facts an individual wishes to conceal, not to gain an edge in transactions, but only to preserve his "constructed public self." See Richard A. Posner, Overcoming Law 539 (1995). Posner's examples include "a newspaper photograph of a person bathing" or a report that one's "daughter was raped and murdered." Id. Even when they are not relevant to any instrumental end, he notes the strong human appetite for concealing such facts about oneself and learning such facts about others. Sec id. at 539-40.

On the other hand, Posner seems to defend an unfettered right to publicize discrediting facts, by which he means true facts that affect one's transactional opportunities. See id. at 539-44. But he does not acknowledge that private acts may be both discrediting and embarrassing, and that the harmful embarrassment that publicity causes the subject can, in principle, exceed the value third parties obtain from the subject's being discredited. The efficient distribution of information need not, therefore, include disclosure of every bit of discrediting data. For a fuller treatment of this issue, see Richard Murphy, Property Rights in Personal Information: An Economic Defense of Privacy, 84 Geo. L.J. (forthcoming July 1996). (99) See supra notes 20-21. (100) If it were necessary for everyone to reach advance agreements with everyone else about what to do upon the discovery of another's secret, the transaction costs would be prohibitive. But blackmail is a two-step process which permits a bargaining opportunity: First. B discovers the information (in the case of opportunistic blackmail, by accident), and second, B releases it. It is not necessary for V or TP to bargain with everyone in the world who might accidentally discover information that they, respectively, wish to conceal or discover. B can negotiate with V or TP after making the discovery, but before making the disclosure decision. (101) See Joseph Isenbergh. Blackmail from A to C, 141 U. Pa. L. Rev. 1905, 1923-25 (1993) (arguing that blackmail, a transaction involving pure information, is particularly susceptible to high informational transaction costs). (102) It takes only one person to "leak" the secret to the world. Even though each V may suffer from disclosure, there are many reasons a particular V might disclose: He may place an unusually high intrinsic value on disclosure, he may be subject to particularly powerful second-party incentives, or he may stand to gain the third-party rewards for being the first to pierce a "conspiracy of silence." The more Vs there are with a common secret, the greater the chances that one will find it desirable to disclose. (103) See Wendy J. Cordon, Truth and Consequences: The Force of Blackmail's Central Case, 141 U. Pa. L. Rev. 1741, 1754 (1993) (arguing by example that a person with information about an embezzling mayor could not practically contact the multitude of voters who might be willing to pay for such information); Posner, supra note 24, at 1822-23 ("Often the benefits of the information will be highly diffuse, being spread across a variety of actual and potential transactors with the blackmail victim, some of whom may not even be identifiable."). In addition, when B stumbles upon a valuable piece of secret information, he is more likely to know who the relevant V is than who the relevant TPs are. He may, of course, know both, but the kind of information that is useful for blackmail identifies V more commonly than it identifies TP. Discovering adultery, for example, usually involves gaining more information about the adulterer than about the adulterer's spouse. Similarly, stumbling upon confidential notes of a psychiatrist is more likely to identify the patient--V--than the persons who would like to know that V is seeing a psychiatrist. Counterexamples clearly exist-discovering evidence of embezzlement might reveal the victim (TP), but not the perpetrator (V). Nonetheless, I believe that adventitiously discovered blackmail information identifies V more often than TP, thus providing another reason why B's costs of dealing with TP are higher than his costs of dealing with V. (104) Thomas Schelling, Micromotives and Macrobehavior 29 (1978). (105) See Isenbergh, supra note 101, at 1923 ("B cannot bargain with [the victim] over the value of the information without revealing some part of it, thereby reducing the amount still undisclosed."); Posner, supra note 24, at 1823 (same). This point is also necessary to complete Kronman's explanation for contract disclosure doctrine. Kronman argues that the law permits contracting parties to withhold information they gain by investment in order to preserve the incentive to make investments. See Kronman, supra note 91, at 9-18 (describing how a duty of disclosure decreases the incentive to invest in information discovery by depriving the knowledgeable party of the private advantage which the information would otherwise afford). If transaction costs were zero, however, parties would contract around whatever disclosure doctrine the law imposed. The party with secret information, for example, would offer to sell the information (for example, that the price of cotton is about to rise or that some land contains valuable minerals) rather than buy the tangible asset (for example, cotton or land) involved. In the end, however, the disclosure rule does matter because of the informational bargaining dilemma. A contracting party in these situations would risk giving his information away in the course of selling it. Thus, we allow the party with information acquired by investment to purchase the tangible assets useful for exploiting that information without first revealing the information to the seller of the assets. (106) Posner makes the same point. See Posner. supra note 24, at 1823. The problem does not arise when the information is proprietary because courts will enforce B's exclusive right to the information. Thus, the analysis does not apply if (1) TP places great value on being able to prove the secret, apart from knowing of its existence, and (2) B has personal property (for example, photographs or financial documents) that provides such proof. In that case, B can do what most buyers do: display his tangible property to TP, but refuse to part with it unless TP meets his price. (107) Reputation can still solve the problem where TP is a repeat buyer of such information or where TP has an agent who acts as a repeat buyer. Tabloid newspapers serve this function for information about celebrities. For example, after receiving a description of the information to be sold, the National Enquirer may offer to pay for the complete story if it subsequently publishes any part of it. People with information might fear that the Enquirer would avoid payment by various forms of strategic behavior. Nevertheless, Bs proceed despite these potential problems because the Enquirer benefits in the long run by maintaining a reputation for dealing fairly with its sources. Consequently, the informational bargaining dilemma may not exist when V is a celebrity. (108) See Posner, supra note 24, at 1842 (finding 124 reported cases involving blackmail prosecutions). (109) This point applies only to nonproprietary information. Where the law creates an enforceable right to exclude others from using information or ideas, one does not risk losing value by describing the information in the course of trying to sell it. Why the law refuses to grant some information the status of property is beyond the scope of this Article. But if the argument in Part II.C is correct, it may be that where the stakes are low, norm-based regulation of information is more cost-effective than the expensive process of enforcing legal rules. (110) If the ATC market failure blocks all "TP transactions, then the question is--for the kind of information that is accidentally discovered-does the universe of Vs value secrecy more or less than the universe of TPs values disclosure? Some people seem to think the answer is obviously "more,. meaning that the universal disclosure resulting from a blackmail ban is worse than universal nondisclosure. But I think that intuition is skewed by the fact that, because blackmail is currently banned, we live in a world in which there are a great many inefficient disclosures and these inefficient disclosures are more likely to come to our attention than inefficient secrets. If one stops taking for granted the efficient disclosures that the ban secures, the overall effect is indeterminate (except for the following textual argument). (111) See supra note 98. (112) See, e.g., J.K. Campbell, Honour, Family and Patronage 272 (1964) ("[S]ince the downfall of one family validates and in some sense improves the status of other families, men attempt by every means of allusive gossip and criticism of conduct to deny each other their pretensions to honour."); Robert Paine, What Is Gossip About? An Alternative Hypothesis, 2 MAN 278,280 (1967) ("[G]ossipers also have rival interests; . . . they gossip . . . to forward and protect their individual interests."); see also Elizabeth Colson, The Makah Indians 201-35 (1953) (discussing the use of gossip in competitive struggles for status). (113) See Richard H. McAdams, Relative Preferences, 102 Yale L.J. 1, 28-48 (1992). (114) See Id. See id. at 55-59. (115) Perhaps even more basic than privacy norms are norms against false gossip. Close-knit groups appear to enforce gossip norms that condemn those who maliciously spread falsehoods about others. See, e.g., Ellickson, supra note 1, at 215 ("Victims who retaliate with gossip are of course tightly constrained by the remedial norm that gossip must be truthful. This general principle is important enough to have earned a spot in the Ten Commandments: 'Thou shalt not bear false witness against thy neighbors.'"); Arthur J. Vidich & Joseph Bensman, Small Town in Mass Society 36 (1968) (noting that the malicious spreader of false gossip is the "most despised person in the community"). Indeed, people use the term gossip pejoratively to refer to precisely these kinds of destructive discussions. See infra notes 121-22 and accompanying text. (116) I do not claim that privacy norms arise in every case where they are needed, or that they work perfectly where they do arise. In general, the sanctioning system will not work if the private gain from the targeted behavior is too high, or if the likelihood of detecting the behavior is too low. Regarding privacy norms, sometimes B will gain a sufficient benefit from embarrassing V-such as intrinsic pleasure from gossiping or a reciprocal favor from TP for sharing the information-that he will willingly bear any likely sanction. Also, sometimes B will be able to disclose anonymously without risk of sanction. If either condition is sufficiently common, a privacy norm may never arise. Even where it does arise, a particular B will violate it if his benefits are still high enough or the risks of detection are still low enough. (117) Examples of the latter include asking how much money someone makes or any of the facts discussed in the next paragraph. (118) See Ferdinand Schoeman, Gossip and Privacy, in Good Gossip, supra note 20, at 72, 79. (119) As a matter of policy, most newspapers refuse to identify rape victims, at least before trial. See Paul Marcus & Tara L. McMahon, Limiting Disclosure of Rape Victims' Identities, 64 S. Cal. L. Rev. 1019, 1021 n.3 (1991). This fact reflects a broader norm against such disclosures. See Naming Names, Newsweek, Apr. 29, 1991, at 26 (reporting that 77% of the public approves of the policy of nondisclosure followed by most news organizations). (120) Similarly, norms against information acquisition facilitate the creation of private spaces necessary for relaxation or self expression, or merely to satisfy an internalized sense of modesty. See Schoeman, supra note 118, at 74 (arguing that social norms restrict access to an individual in certain domains in order to promote individuality, private life, and the integrity of various spheres of life). (121) See Edith B. Celles, Gossip: An Eighteenth-Century Case, 22 J. Soc. Hist. 667, 667 (1989) ("[E]vidence from many different sources implies that in every age people have gossiped, and for as long as literature has referenced gossip, its reputation has remained negative."); see also Spacks,, supra note 20, at 24-26 (discussing the "reputation" of gossip). (123) See John Sabini & Maury Silver, Moralities of Everyday Life 91-93 (1982) (claiming that the pejorative term "'gossip' highlights the idleness of talk," its irrelevance to any plan of action); Nicholas Emler, Gossip, Reputation, and Social Adaptation, in Good Gossip, supra note 20, at 117, 120 (noting that "the term 'gossip' is often used in common parlance not as a neutral description . . . but as a moral judgment"). (125) See Ross Sneyd, When Solzhenitsyn Departs, Life Won't Be the Same for Vermont Town, Chi. Trib.,.,June 11, 1993, [sections] I (News), at 8 (noting that "many [residents] are doggeddefenders [sic] of the family's privacy."). The owner of one general store kept a sign warning tourists, reporters, and other passersby: "No Directions to the Solzhenitsyn Home." One dairy farmer who lived near Solzhenitsyn explained: "This is the way Vermonters are .... If you want to mingle, you mingle. If you don't, you don't." Id. Admittedly, this example deals with how a group keeps information secret from outsiders rather than how a group controls distribution of information within itself. But in either case, I only allude to the fact that privacy norms may be group-specific. For a discussion of "conspiracies of silence" and their tendency to be inefficient, see infra note 137. (124) See supra part I.A. (125) Indeed, the argument that follows may understate the comparison in assuming that comparable disclosure norms even exist. If B's average expected gain from preserving V,s secret is sufficiently high, a norm of disclosure will never arise. (126) The same comparison holds between privacy norms and norms against blackmail. One might imagine that, if a group can enforce privacy norms, it can enforce antiblackmail norms as well. But the existence of privacy norms, some unsupported by law, does not imply that groups could entirely suppress blackmail without state sanctions. Even privacy norms are underenforced. See supra note 116. Absent a legal prohibition, antiblackmail norms would be enforced even less than privacy norms. Antiblackmail norms are essentially the same as disclosure norms, refraining from blackmail being the first step toward disclosure. Norms against blackmail would therefore suffer the same weaknesses: (1) the violation of an antiblackmail norm tends to be a complete secret, so the probability of detection is lower than it is for violating privacy norms; (2) given asymmetric transaction costs, the bribe V offers B to violate an antiblackmail norm (the blackmail price) will, on average, be greater than the bribe TP pays B to violate a privacy norm. Thus, the existence of privacy norms unsupported by law does not imply that groups could fully suppress blackmail without state sanctions. Cf. supra notes 24, 72.

That even privacy norms are underenforced, however, raises another question: Why does the state not enforce all efficient privacy norms? I will not attempt to answer this question fully; it is part of a much broader debate over norm governance. I only note the trivial point that law should not incorporate all norms because of the administrative costs to enforcing law. The costs of legal enforcement will exceed the benefits if the stakes involved are low or the norm is too complex and too dependent on local knowledge. See, e.g., Bernstein, supra note 8, at 1785-95 (noting that, because of litigation costs and judicial error, some contracting parties prefer that their business norms remain legally unenforceable); Saul Levmore, The Anonymity Tool, 144 U. Pa. L. Rev. 2191, 2216 (1996) ("We might think of intermediation as one of many mechanisms that thrive in informal contexts, but that are not easily formalized."); Edward B. Rock 8c Michael L. Wachter, The Enforceability of Norms and the Employment Relationship, 144 U. Pa. L. Rev. 1913, 1932-38 (1996) (arguing that law should not incorporate the norm of "for cause" dismissals because of problems of local knowledge). The state does create certain legal obligations of confidentiality, but does not obligate privacy in every situation where a privacy norm applies because such norms can be very complex and very dependent on local knowledge. See supra notes 118-19 and accompanying text. For example, Robert Post characterizes tort liability for privacy intrusions as incorporating community norms of "civility." See Robert C. Post, The Social Foundations of Privacy: Community and Self in the Common Law Tort, 77 Cal. L. Rev. 957, 1004-08 (1989). But because of the difficulties in translating "complex, tacit, and contextual territorial principles" into "the relatively clear, explicit, and precise elements of a formal cause of action," the tort incorporates only the "most important" community norms. Id. at 975. (127) My argument obviously holds constant the sanction cost the group imposes for violating either norm. In theory, the higher benefit B receives for violating a disclosure norm and the lower probability his violation is detected could be offset by a rise in the sanction cost B bears if his disclosure norm violation is detected. Absent some theory of perfect homeostasis of norm enforcement, however, this outcome is excessively optimistic. A more realistic assumption is that sanctioning is costly and scarce and that a group cannot raise the sanction intensity for one norm without lowering the sanction intensity for other norms. If blackmail were legal, perhaps a group would reallocate its norm enforcement capital to raise the sanction cost for violating disclosure norms above the sanction cost for violating privacy norms. But given some costs of doing so, the group would probably not completely offset the change in the other enforcement variables. (128) Note that many such statutes forbid or create liability for disclosure by institutions that might not be subject to privacy norms. See, e.g., Privacy Act of 1974, 5 U.S.C. [sections] 522a (1994) (setting requirements for the disclosure of medical, educational, criminal, and financial information held by agencies of the government); Family Education Rights and Privacy Act, 20 U.S.C. [sections] 1232g (1994) (protecting privacy of student records); Video Privacy Protection Act of 1988, 18 U.S.C. [sections] 2710 (1994) (preventing disclosure of information about consumers by video tape service providers). (129) Even with a ban on blackmail, the problem of excessive secrecy may be so great that the state will sometimes need to require disclosure or provide rewards for disclosure. (130) There is yet another asymmetry favoring the blackmail ban. Even where the ban and the lawful blackmail market distribute information identically, there is still a difference of whether one has to pay for secrecy or disclosure. If blackmail is legal, Bs will threaten to breach privacy norms unless paid their blackmail price. Vs will sometimes have to pay for efficient levels of privacy they would otherwise enjoy for free. The harm in permitting B to extract payments from V by threatening to violate an efficient norm is analogous to the harm in permitting B to extract payment from V by threatening to violate an efficient criminal or tort rule. Such threats cause wasteful defensive investment of the sort that would occur if the law did not sanction the actual commission of the crimes and torts. If the privacy norm is efficient, permitting people to extract payments by threatening to breach the norm is inefficient. Note the different result for the reverse situation: If we ban blackmail, TPs will not have to pay for efficient disclosures. Transaction costs make it very difficult for B to extract payment from TP. Threatening not to gossip about the subject unless paid invokes the informational bargaining dilemma, where B cannot fully describe the information he is selling without giving it away. Thus, where disclosure is efficient, B will not be restrained by privacy norms, and he will tend to gossip. TP will not have to pay to learn what others adventitiously discover.

One might object that when a norm counsels against disclosure, B's threat to disclose is not credible because B will suffer a cost by disclosing. But some of these threats will be credible. For example, B may threaten to avoid the norm sanction by disclosing the secret anonymously. Or, if V can use B's threat to identify him as the likely source of an anonymous disclosure, B may conceal his identity from V, anonymously threatening to disclose. In addition, V must consider the possibility that B has miscalculated because, for example, B wrongfully believes that no privacy norm exists or that, for some reason, his violation will be excused. When the stakes are high, as will often be the case with private information, even a low risk that B has miscalculated may make his threat credible. See Daniel Ellsberg, The Theory and Practice of BlackMail, in Bargaining: Formal Theories of Negotiation 343 (Oran R. Young ed., 1975) (explaining the success of robbery threats where the victim's cooperation is required). (131) See supra note 33. (132) This is particularly true because, from V's perspective, the "critical information" is not the substance of his embarrassing behavior or characteristic, but information identifying him with that behavior or characteristic. B will rarely enjoy passing on "identity information" to TP if neither he nor TP knows V. Suppose, for example, that B accidentally discovers that a stranger is having an adulterous affair, is psychotic, or suffers from a sexual dysfunction. The details of the matter or its discovery may be interesting and pleasurable to recount. But everyone knows that adultery, mental disease, and sexual dysfunction exist; unless the listeners know the subject, his name is neither relevant nor amusing. Including unimportant details in a story diminishes the pleasure of the recipients and, hence, the teller's pleasure as well. (Celebrities may present an exception because people like to gossip about celebrities by name whether or not they personally know them.) (133) Ellickson, supra note 1, at 181; see also Laing, supra note 17, at 38 ("The more highly organized or exclusive a group, the more gossip there is within it."): Noon & Delbridge, supra note 17, at 27 (suggesting that "the extent of gossip may reveal how tightly knit a community is.... [T]he better integrated the group, the more free the gossip."). (134) In addition, the opportunity for V accidentally discovering information about B probably increases as the interaction between V and B increases. Thus, the people most likely to stumble across damaging or embarrassing secrets about V are members of V's social groups. (135) See notes 94, 97 and accompanying text. The ban also allows the parties to reach the same outcome-silence-without incurring bargaining costs. (136) Because we are discussing opportunistic blackmail only, I assume that blackmail will probably decrease the cost of maintaining secrets. Permitting only opportunistic blackmail would not raise the investment in information, so it would not affect the probability of discovery, but V would pay less in blackmail than the costs he would expect to bear upon full discovery. See supra notes 28-32 and accompanying text. (137) Another complication is the existence of "hybrid" cases, in which norms affect only two of the three parties to a possible blackmail transaction. For example, B may belong to a group with TP but not V; B may belong to a group with V but not TP; or B may belong to a group with V and a different group with TP. The second and third of these cases may give rise to a very important phenomenon-the "conspiracy of silence." Here, a group enforces an informational norm forbidding disclosure of negative information about members to nonmembers. Illustrations include cooperative silence among police, doctors, and crime "families." Such norms are likely to facilitate group activities that harm the interests of those outside the group, as where they prevent sanctions for crimes or torts. Thus, conspiracies of silence are frequently inefficient. (Magicians are an interesting exception; their tricks must be secret to be entertaining.)

The ban on opportunistic blackmail probably undermines conspiracies of silence. With a conspiracy, we have a group of people who know each other's secrets and who benefit, as a whole, from continued silence. The incentive structure is similar to the original prisoner's dilemma, except that repeated future interaction makes silence the dominant strategy. For this reason, the threat by one conspiracy member to divulge the secret of another is usually not credible. The group struggles to control membership and sanctions in a way that ensures such threats remain less than credible. But occasionally, a member will convince the group that he has nothing to lose from disclosing, perhaps because he has developed greater loyalty to another group or because his own secrets have already been discovered. At this point, a blackmail transaction will permit the group to maintain a conspiracy of silence where it otherwise would not. The blackmail ban tends to break down the conspiracy of silence in these cases by facilitating the willingness of these dissident B's to disclose. The blackmail price paid by the group seeking secrecy will be less than the sanction the group would pay were the information disclosed. See supra notes 28-32 and accompanying text. So blackmail would decrease the sanction ultimately imposed on conspiracies. And for opportunistic blackmail, there is no compensating increase in the probability of discoveries; if one invests to gain information for blackmail purposes, then one is no longer engaging in adventitious blackmail. (Nor is commercial blackmail necessary to undermine such conspiracies because the state or other organizations can offer more calibrated rewards, as the state does when it compensates and protects "informants" and "whistleblowers." Cf. Shavell, supra note 28, at 1899-1900 (arguing that a reward policy is superior to allowing blackmail of criminals).) (138) Of course, if one is confident that permitting opportunistic blackmail lowers the price of maintaining secrets, there would be an efficiency gain from permitting blackmail only for the kind of secrets we think it is efficient for people to keep concealed. But imagine the complexity of an exception permitting only (1) opportunistic blackmail (2) in which the parties are not members of the same close-knit group (3) but would still gossip (4) about matters that are best kept private. Although it is best to avoid relying on administrative costs too easily, see supra note 93, at this point they provide a rather obvious basis for preferring a simpler rule prohibiting all blackmail.

Richard H. McAdams, Associate Professor of Law, Chicago-KeDt College of Law. I thank Anita Bernstein, Dennis Chong, Wendy Cordon, Rick Hasen, Hal Krent, Jim Lindgren, Anna Marshall, Eric Posner, Dan Steward, and the participants at workshops at North,western University and the University of Florida for their insightful comments on earlier drafts. I also thank Cholley Kuhaneck for providing research assistance. My work on this Article was supported by the Marshall D. Ewell Research Fund.
COPYRIGHT 1996 University of Pennsylvania, Law School
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1996 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Symposium: Law, Economics, & Norms
Author:McAdams, Richard H.
Publication:University of Pennsylvania Law Review
Date:May 1, 1996
Words:26308
Previous Article:The anonymity tool.
Next Article:The Song of Deborah: a legal-economic analysis.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters