Encyclopedia of Law and Economics
Show Less

Encyclopedia of Law and Economics

Edited by Gerrit De Geest

The second, expanded edition of the acclaimed Encyclopedia represents a major update of the most authoritative reference work in the field of law and economics and the nine print volumes are now released online as a single integrated product.The Encyclopedia provides balanced and comprehensive coverage of the major domain in law and economics, including: criminal law, regulation, property law, contract law, tort law, labor and employment law, antitrust law, procedural law, and the production of legal rules. Each theme or volume is overseen by a leading scholar and each of the 166 entries is prepared by an expert in the field, providing an in-depth and authoritative overview of the individual topic, combined with an exhaustive bibliography, allowing users to access and filter the entire corpus of literature in law and economics.As with the print edition, the Encyclopedia is unique in serving both as an entry point and a platform for advanced research. The online edition is enhanced with Elgaronline’s powerful search tools, facilitating the search for key terms across the entire Encyclopedia, whilst the browse function allows users to move seamlessly between the volumes. These elements combine to create a powerful research tool for any researcher or scholar in the field of law and economics.
Show Summary Details
This content is available to you

Chapter 14: Self-incrimination

Alex Stein

[In: Volume 8, Chris William Sanchirico (ed) Procedural Law and Economics]

1. Introduction

The Fifth Amendment to the United States Constitution provides that “nor shall [a person] be compelled in any criminal case to be a witness against himself.”

This provision is widely known as the privilege against self-incrimination or the right to silence. The right to silence incorporates four basic rules. As a general matter, a person who receives a subpoena or other lawful request to provide information to an authorized tribunal or agency, such as a court or the police, but refuses to do so, is guilty of contempt or a similar crime punishable by fine or imprisonment. The person, however, is exempted from the duty to comply with such a request when his compliance might reveal information exposing him to a prospect of criminal prosecution and conviction. Second, factfinders may not draw any adverse inferences from a criminal defendant’s refusal to testify in his defense or answer questions during police interrogation. Third, when the police or other law-enforcement agency elicits an involuntary confession from a suspect, the confession cannot be admitted into evidence and the trial judge must suppress it. Coercive interrogation of a suspect renders the ensuing confession involuntary and inadmissible.1 Furthermore, a suspect’s confession will be deemed involuntary as a matter of law – and, consequently, inadmissible – when the police deprive him of his Miranda rights at custodial interrogation. Under Miranda, the police must tell the suspect at the beginning of his interrogation that he is entitled to remain silent; that anything he will say might be used as evidence against him at his criminal trial; and that he is entitled to consult an attorney (at his own expense or at the government’s expense, if the suspect is poor) and have that attorney present at the interrogation.2 Finally, criminal defendants and suspects can waive each of the above entitlements. If the waiver is a product of a voluntary and informed decision, the court will recognize it as effective.

p. 367In what follows, I survey these rules and their underlying economic justifications.

2. The Utility of the Right to Silence

2.1 The Right to Silence as an Anti-pooling Device

Economic analysis of the right to silence focuses on the right’s social costs and benefits. If those benefits exceed the costs, the right will be justified. The right’s costs and benefits crucially depend on how it affects the outcomes of criminal investigations and trials. Specifically, those costs and benefits depend on how the right affects the incidence of false positives (erroneous convictions of factually innocent defendants) and false negatives (erroneous exonerations of factually guilty defendants). The right to silence will be justified if it reduces the total social cost of false positives and false negatives in comparison with an alternative legal regime that does not recognize the right.3

False positives and false negatives are the consequences of asymmetrical information. A criminal defendant normally knows for certain whether he “did it” or not. The police, prosecution, and courts have no such knowledge. False positives and false negatives are brought about by defendants’ choices between staying silent, confessing to the crime and denying the accusations, and by the actions of other participants in the criminal process – the police, prosecution, and courts – that respond to those choices (Seidmann and Stein 2000). These strategic interactions merit detailed analysis.

The first thing to know about the right to silence is that it plays no significant role in cases in which the outcome of the defendant’s criminal trial is virtually certain. The right has no effect on a case in which inculpatory evidence is overwhelmingly strong. By the same token, it has virtually no effect on a case featuring weak inculpatory evidence. In the former category of cases, both guilty and innocent defendants face a serious prospect of conviction. The right to silence cannot change this prospect in either direction. For a defendant who faces overwhelming inculpatory evidence, making a confession followed by a guilty plea would normally be the best strategy. This strategy might secure a sentence reduction, and it also would allow the defendant not to expend money and effort on litigating a hopeless case.

p. 368Defendants facing weak inculpatory evidence will likely be exonerated, regardless of their factual guilt. As in all other cases, an innocent defendant’s best strategy in a weak-evidence case is testifying and telling the truth. The prosecution’s weak evidence would fail to refute his true exculpatory testimony. The factfinders consequently would have to acquit the defendant. A guilty defendant, in contrast, must always choose between giving a perjured self-exonerating account of the events, staying silent, and confessing. Making a confession followed by a guilty plea is only attractive when the prosecution offers the defendant a favorable plea bargain. Absent such an offer, a guilty defendant must choose between lying and remaining silent. The lying strategy is risky: an uncovered lie would reveal the defendant’s “guilty conscience,” which would practically guarantee his conviction.

Whether a guilty defendant should remain silent depends on the legal regime. When the legal regime does not allow factfinders to draw adverse inferences from the defendant’s silence at interrogation and trial, the defendant’s best call is to remain silent. When adverse inferences are allowed, the defendant’s choice between silence and lies would depend on how strong those inferences are. If those inferences merely indicate the defendant’s possible involvement in the crime – and thus function merely as corroborative evidence – the defendant should stay silent. The prosecution would then fail to prove his guilt beyond a reasonable doubt, and the factfinders would have to acquit him. However, if factfinders always (or predominantly) associate silence with guilt, the defendant would be better off lying. The right to silence therefore has no effect on defendants with exceedingly high payoffs for lying or confessing. For those defendants, the right is essentially irrelevant (Stein 2008).

The right to silence, however, plays a significant role in factually complex (“intermediate”) cases, in which the inculpatory evidence is fairly (but not overwhelmingly) strong. These cases present the most acute problem of asymmetrical information. Factfinders know that some defendants are guilty and some innocent, but cannot tell who is guilty and who is innocent. Virtually every defendant knows whether he committed the crime that the prosecution accuses him of. His private knowledge, however, does not turn into public information that factfinders can verify and trust. In the absence of a special incentive to plead guilty (a particularly attractive plea-bargain offer), a guilty defendant will deny the accusations and plead not guilty. Under the regime that allows factfinders to draw adverse inferences from the defendant’s silence, guilty defendants will falsely testify about their innocence. Innocent defendants will do the same, but without lying. Factfinders consequently will proceed on the assumption that some self-exonerating accounts are true and some false.

p. 369This assumption necessarily reduces the probability of all self-exonerating accounts. As a result, an innocent defendant who cannot corroborate his exculpatory testimony by credible evidence suffers a credibility reduction. This reduction increases the probability of the prosecution’s case – an increase that helps the prosecution prove the defendant’s guilt “beyond a reasonable” doubt. When that happens, factfinders convict an innocent defendant, which means that, by lying, a guilty defendant imposes a harmful externality on innocent defendants (and society at large) (Seidmann and Stein 2000). Bentham’s famous utilitarian analysis of the right to silence, maintaining that the right only helps guilty criminals to escape conviction, failed to notice this externality. This analysis, therefore, is seriously flawed. The present-day supporters of “crime control” – who claim, similarly to Bentham, that the right only helps guilty criminals to escape conviction – have also overlooked it. Contrary to the crime-control view, abolition of the right to silence would not induce guilty criminals to switch from silence to confessions.4 Rather, it would p. 370induce them to switch from silence to self-exonerating lies (Seidmann and Stein 2000, at 499–502).

The right to silence gives guilty criminals an attractive alternative to lying. Because a lie can be discovered and because its discovery will likely lead to the liar’s conviction, for at least some criminals silence constitutes a better option. Those criminals would consequently prefer silence to lying. The externality that they otherwise would impose upon innocents (the pernicious pooling effect) will thus be eliminated. As a result, fewer innocent defendants will be convicted than under a regime in which the right to silence does not exist. This outcome, however, will be achieved at a price: some criminals, who otherwise would implicate themselves by lies that could be uncovered, will escape conviction by exercising the right to silence (Seidmann and Stein 2000; Stein 2008).

2.2 The Doctrinal Fit

This anti-pooling rationale is particularly useful as an explanatory tool. It explains and justifies the entire set of rules that derive from the right to silence. These rules hold that:

  • The right to silence protects defendants throughout the entire criminal process, which includes interrogation, trial and sentencing hearings.

  • The right to silence protects defendants only against compelled disclosure of “testimonial,” but not “physical,” evidence.

  • The right to silence (in the form of the privilege against adverse inferences from silence) does not apply in civil trials.

  • Nor does it extend to testimony that may lead to the witness’s conviction overseas.

  • p. 371The right to silence can be set aside when the police attend an ongoing emergency.

  • The right’s protection against compelled self-incrimination is given only to people, but not to corporate entities.5

The remainder of this section will explain and justify each of the above rules in the order presented.

2.2.1 The right to silence as applying to every phase of the criminal process

By allowing non-confessing criminals to substitute silence for false self-exonerating statements, the right to silence protects innocent defendants from being pooled with criminals. At the same time, it allows criminals not to implicate themselves by potentially incriminating lies. The right to silence therefore increases the rate of erroneous exonerations in factually complex cases. It does not merely help innocents to achieve deserved acquittals. It also helps criminals to escape conviction. The exoneration prospect, indeed, is the prize that a criminal receives for helping the legal system to cleanse the pool of self-exonerating statements. The resulting exonerations of the guilty consequently determine the right’s social cost. For obvious reasons, the legal system is interested in reducing this cost without sacrificing the right’s anti-pooling benefit. This desire explains the right’s limits discussed below in subsections 2.2.2, 2.2.3, 2.2.4, 2.2.5 and 2.2.6. The right’s cost, however, is also a reason for questioning its broad application throughout all phases of the criminal process. Presently, the right to silence applies with full force at interrogations, trials, and sentencing hearings. Is there a good reason for that? Would the right’s removal from any of those phases be socially beneficial?

The right to silence cannot be removed from suspects interrogated by the police: its removal will be too costly. If factfinders draw adverse inferences from the defendant’s silence at interrogation, guilty suspects would have an incentive to lie to the police. Subsequently, those suspects would have to repeat their lies while testifying in court (if they do not do so, the prosecution would often be able to use their false statements to the police as evidence of guilt).

These suspects’ false statements will consequently pool with the uncorroborated true testimonies given by innocent defendants. Facing this pooling, factfinders will have to reduce the probability of truthfulness that attaches to all such self-exonerating testimonies.

The right to silence also cannot easily be removed from the criminal trial. p. 372For reasons already given, this removal would motivate guilty defendants to pool with innocents by falsely testifying in their defense – a pooling that would cause factfinders to discount the credibility of every defendant’s testimony. Apart from creating this pernicious effect, the right’s unavailability at the trial stage would motivate guilty suspects to lie to the police as well. Those lies would often be a guilty defendant’s strategic necessity.

Anticipating the prosecution’s attempt to rebut his testimony at a forthcoming trial, a guilty suspect would often need to give a prior consistent statement that will bolster his credibility as a witness (Seidmann and Stein 2000, at 489–95).

After pleading guilty, a defendant retains the right to silence at his sentencing hearing. In determining sentencing facts, the judge cannot draw adverse inferences from the defendant’s failure to testify at the hearing. This rule, too, has a compelling anti-pooling rationale. Removal of the right to silence from sentencing hearings would induce guilty defendants to plead not guilty. Some of those defendants would remain silent at their trials and enjoy the pre-conviction protection against adverse inferences. Others would falsely testify to their innocence and adversely affect innocent defendants by impugning the credibility of their truthful self-exonerating accounts. Neither of those scenarios is beneficial to society (Seidmann and Stein 2000, at 495–8).

2.2.2 The right to silence as restricted to “testimonial” evidence

The right to silence only protects defendants from compelled production of “testimonial” evidence. The right does not extend to “physical” evidence, which includes writings that already exist (as opposed to writings that a suspect or a criminal defendant might be asked to generate).

This limitation squarely aligns with the anti-pooling rationale. Under this rationale, the right to silence should only be recognized when guilty defendants need inducement for avoiding pooling with the innocent. The right should only extend to evidence that can create this pooling effect. Such externality-laden evidence reduces the credibility of self-exonerating evidence tendered by the innocent. Any such evidence therefore should be considered “testimonial” for purposes of the right to silence, and defendants should not be required to produce it. All other evidence should be categorized as “physical”; the right to silence should not protect defendants against compelled disclosure of such evidence. Factfinders, consequently, should be authorized to draw adverse inferences against a defendant who refuses to provide “non-testimonial” or “physical” evidence (Seidmann and Stein 2000, at 475–81).

Utterances and their non-verbal equivalents – for example, sign language and a person’s nodding of her head for a “yes” – clearly fall into the p. 373“testimonial” category. But evidence would also classify as “testimonial” in any case in which its producer can shape its content and meaning. This evidence-shaping ability makes the evidence externality-laden. For example, a handwriting sample that a suspect produces at the police station is “testimonial” because a guilty suspect might replicate an innocent person’s handwriting.6 By contrast, handwritings that already exist fall into the “physical” evidence category because their contents and meanings do not depend on the defendant’s thought processes; a qualified expert (and sometimes a nonexpert witness as well) can authenticate such samples.

By the same token, the right to silence protects the defendant against compelled production of a document that involves an implicit acknowledgment of his possession of the document and of the document’s authenticity and relevancy to the trial. To obtain a potentially inculpatory document (“physical” evidence for purposes of the Fifth Amendment), the government must lawfully seize it from the defendant. To compel the defendant to produce the document, it must first guarantee that the defendant’s act of production – an implicit testimonial activity – will not be used as evidence against him. Absent this “use immunity,” the defendant will be free not to produce the document. Furthermore, the government cannot compel a defendant to assemble documents for its criminal investigations even when it guarantees use immunity. Assembling documents is identical to testifying about the documents’ nature and contents.7

Both production and assembly of documents are externality-laden because their results depend on the producer’s or assembler’s choice and opportunity to manipulate. A guilty defendant will produce only those documents that can help him establish his innocence. Aware of this self-serving motivation, factfinders will reduce the credibility of any assembly and production of documents by a criminal defendant and of any claim that defendants can make about documents’ existence, authenticity, and custody. As a result of this undeserved credibility-reduction, some innocents might be found guilty (Seidmann and Stein 2000).

2.2.3 The right to silence as confined to criminal trials

The rule against adverse inference from the defendant’s silence only applies in criminal p. 374trials. In civil cases and other non-criminal proceedings, such inferences are generally allowed.

The anti-pooling rationale fully justifies this limitation of the right to silence. The pooling problem that the right attenuates does not exist in civil and other non-criminal proceedings because those proceedings do not involve innocents who face the possibility of wrongful conviction and punishment. To be sure, the unavailability of the right motivates liars to pool with truth-tellers in those proceedings as well, but this pooling occurs outside the machinery of criminal justice. The legal system consequently need not sacrifice probative evidence in order to prevent or mitigate this pooling effect. There is no good reason for doing so in civil cases, where the cost of false positives and false negatives is roughly the same.8

2.2.4 The right to silence as confined to same-sovereign prosecutions

A witness in a state proceeding can invoke the privilege against self-incrimination out of concern regarding a federal prosecution, and vice versa. By the same token, a witness in a state proceeding can successfully claim the privilege by referring to a prosecution in another state. The privilege, however, will not apply when the witness’s testimony (or disclosure of other protected information) exposes him to the prospect of conviction outside the United States.

The anti-pooling rationale justifies this same-sovereign limitation as well. The limitation generates no pernicious pooling inside the American criminal justice system. When a criminal tried in another country chooses to lie, his lies do not increase the risk of wrongful conviction for innocent defendants in the United States. The externality that his lies generate occurs overseas. Consequently, there is no need to eliminate this externality by upholding the right to silence that taxes the American system. The same-sovereign limitation helps generate probative evidence for American courts and law-enforcement agencies, and there is no good economic reason to forego this benefit in order to protect foreign innocents. Those innocents ought to be taken care of by their own legal systems.9

2.2.5 The emergency exception to the right to silence

Under the emergency exception to the right to silence, a self-incriminating statement that the police obtain from a suspect while attending an ongoing crime-related p. 375emergency is admissible as evidence at the suspect’s subsequent trial regardless of whether the suspect received the Miranda warnings.10

The right’s functioning as an anti-pooling device justifies this exception. Statements that this exception makes admissible are invariably inculpatory. As such, they never pool with self-exonerating accounts of innocent defendants.

2.2.6 The right to silence as belonging to persons, not corporations

The right to silence protects only natural persons, as opposed to corporations. Nor does it extend to a corporate agent or employee who is required under the color of law to provide documents or other information tending to incriminate the corporation. A corporate agent or employee can only claim the right in his personal capacity; and even this personal entitlement is qualified by the “collective entity” rule. Under this rule, a person’s assumption of a corporate job entails a duty to produce corporate documents regardless of the self-incriminating consequences to the person. This special rule intensifies deterrence against corporate crime (Bierschbach and Stein 2005).

Forcing a corporate insider to testify against her corporation induces the insider to lie. This perjured testimony pools with true testimonies of insiders of other corporations that face criminal accusations. This pooling increases the risk of undeserved conviction for innocent corporations. But the repercussions of such convictions for a natural person – a stockholder, a director or an employee – are strictly pecuniary, as in civil cases in which false negatives are as harmful as false positives. Consequently the right to silence does not apply.

The “collective entity” rule, therefore, is the only serious departure from the right to silence. A corporate insider must comply with a legal requirement to produce documents even when the act of production implicitly acknowledges some fact that might incriminate the insider. Unlike regular defendants, the insider will not be entitled to “use immunity” under such circumstances. As a result, a guilty insider’s production of innocent-looking documents will pool with an innocent insider’s production of innocent documents. This pooling will cause factfinders to reduce the credibility of innocent insiders to the detriment of those insiders.

This credibility reduction and the consequent increase in the rate of erroneous convictions are socially undesirable. The “collective entity” rule, however, can still be justified as a means of increasing the law-enforcers’ access to corporate documents. This access facilitates the p. 376enforcement of corporate liability for fraud and other illicit activities that often go undetected.

2.3 Variations

The anti-pooling rationale of the right to silence has four important variations. An attempt has been made to estimate the right’s effect on social welfare (Seidmann 2005). This attempt produced a formal model in which factfinders receive negative payoffs for each erroneous conviction and acquittal. To align the factfinders’ preferences with society’s welfare, the payoffs are set to represent the socially accepted tradeoff between erroneous acquittals and erroneous convictions. Correspondingly, factfinders receive –D when they convict an innocent defendant and −(1−D) when they acquit a guilty defendant. D is the probability threshold for convictions: when the probability of a defendant’s guilt equals D, the factfinders are indifferent between acquitting and convicting him. Under this model, the right to silence reduces the rate of erroneous convictions as a by-product of raising the incidence of erroneous acquittals. The right consequently reduces social welfare by systematically suppressing probative inculpatory evidence: the defendant’s unwillingness to speak to the police or to testify in his defense. This suppression causes factfinders to acquit defendants whose real probability of guilt is greater than D and who, in all likelihood, are guilty as charged.

The big question, however, is whether the avoided convictions of innocent defendants can improve social welfare substantially enough to offset the harm caused by the acquittals of the guilty. Consider a setup in which the right to silence is not available and the pooling problem is particularly acute. As a result, factfinders discount the credibility of all uncorroborated self-exonerating testimonies. This discounting makes the conviction of an innocent defendant f times more likely than under the previous regime. To tackle this problem, the system’s designer modifies the payoffs for factfinders. From now on, factfinders will pay fD, instead of D, for every conviction of the innocent, and (1 – fD), instead of (1 – D), for every acquittal of the guilty. Will those factfinders agree to switch to a legal regime in which their payoffs are set as before, but the right to silence is available? The answer to this question crucially depends on f. This figure may heighten the level of proof for convictions to a degree that would hardly allow the factfinders to convict any criminal. The right to silence, therefore, might bring along an important sorting benefit that will legitimize a socially beneficial reduction of the criminal standard of proof. This insight verifies an observation that the right to silence and the criminal proof standard are complementary policy tools (Seidmann and Stein 2000).

p. 377Second, in some settings, the right to silence can benefit innocent defendants directly. Because inculpatory evidence may be inaccurate, it may indicate that the defendant is guilty even when he is innocent. The defendant’s true self-exonerating story consequently may appear false, thus further increasing the probability of his guilt in the factfinders’ eyes. For example, the defendant may have a completely true alibi contradicted by a number of perjurous or mistaken witnesses that have credible appearance. For such defendants, the best trial strategy is silence. A rule that prohibits adverse inferences from silence therefore can help innocent defendants by preventing factfinders’ error (Leshem 2010).

Third, the right to silence may be combined with stringent disclosure requirements from the prosecution. These requirements will further motivate innocent defendants to disclose exculpatory evidence, while their guilty counterparts still prefer silence or lies. This dynamic will increase social welfare by bolstering the separation between guilty and innocent defendants (Mialon 2005).

Finally, there is a good economic reason for extending the privilege against adverse inferences to some civil cases. The unavailability of this privilege under extant law motivates civil defendants to avoid adverse inferences by searching for exonerating evidence. If the evidence that a defendant finds is likely to exonerate him, the defendant would happily adduce it. This evidence would forestall the adverse inference and help the defendant defeat the lawsuit. On the other hand, if the evidence identifies the defendant as potentially (or actually) liable, the defendant would simply suppress it. Theoretically, he must comply with the discovery rules and let the plaintiff have this evidence. But the actual enforcement of those rules is far from perfect, and the defendant would exploit this shortcoming. He would take advantage of the plaintiff’s unawareness of his discovery and possession of the evidence unfavorable to his case.

The defendant will search for favorable evidence whenever its expected value is greater than the cost of the search. As I already indicated, the imperfect discovery regime allows the defendant to ignore the prospect of finding unfavorable evidence. Because the defendant can hide such evidence, his expected loss from finding it equals (roughly) zero. The defendant’s search for evidence consequently has little or no effect on his primary activity and its risks of harming another person. Evidence that the defendant finds and subsequently hides or adduces may affect the outcome of the litigation, but this effect is merely a transfer of wealth from one party to another, which is of no social value. Hence, the defendant’s expenditure on the search is privately beneficial, but socially wasteful. The defendant’s incentive to carry out a socially inefficient search for evidence thus needs to be reduced. To this end, the lawmaker can interpose a rule blocking p. 378adverse inferences from a defendant’s failure to adduce exonerating evidence (Wickelgren 2010).11

3. The Economics of Confessions

3.1 The Meaning of “Voluntariness”

Defendants confessing to a crime substantially increase their probability of being convicted and punished. With all other things being equal, confession reduces the confessor’s welfare. The magnitude of this reduction – the expected harm from confessing – equals the increase in the confessor’s probability of being convicted multiplied by the punishment for the underlying crime.

But for most defendants, “all other things” are virtually never equal. First, some defendants – guilty and innocent – may be facing strong inculpatory evidence and a correspondingly high probability of conviction. For them, confessing to the crime and subsequently entering a guilty plea would often be a better call than spending money and effort on a hopeless trial.

Second, the punishment for convicted defendants who did not confess and plead guilty may be set much higher than the punishment for confessors. The increase in the expected punishment may induce some defendants – guilty and innocent alike – to confess and plead guilty even when the inculpatory evidence is not strong.

Third, the law may separately punish defendants for staying silent and for lying during police interrogation or in court. For defendants whose self-exonerating stories may be found false – rightly or wrongly – this special penalty would have the same effect as an increased punishment for non-confessors. For guilty defendants who consider remaining silent, the effect of the penalty for silence would be different. Facing this penalty, some guilty defendants might decide to lie in their defense. By doing so, they would pool their false exculpatory statements with true self-exonerating accounts tendered by innocent defendants. Factfinders consequently would discount the probability of true exculpatory statements that have no corroboration. This socially deleterious dynamic was explained in Section 2.

Finally, confessing to a crime may remove physical and psychic pressures that the police may exert upon suspects, both guilty and innocent. Those pressures may be strong enough to elicit a confession from both types of defendant.

p. 379The upshot is that confessions are never given for free.12 They are triggered by fear and favor. Fears instilled and favors promised define the defendant’s benefit from confessing. A rational defendant confesses to a crime only in exchange for a benefit: a sentence reduction, a saving of trial expenses, and sometimes removal of a threat that comes from the underworld or a state agent. Some of those benefits motivate guilty defendants to confess and plead guilty. Other benefits extract false confessions from innocent defendants. Neither benefit elicits a confession that can be considered factually “voluntary.”

Factual voluntariness, indeed, makes no sense at all because criminal suspects and defendants are not free from interrogations, trials, custody, and bail restrictions. These mechanisms of criminal justice situate every defendant under fears and favors that elicit confessions. Those fears and favors make virtually any confession factually involuntary.13 Courts in the United States therefore did not adopt factual voluntariness as a criterion for the admissibility of confessions. The prevalent voluntariness criterion for confessions’ admissibility is normative, rather than factual.

The courts’ voluntariness jurisprudence tackles two fundamental problems. The first problem is pooling of the guilty and the innocent. To resolve this problem, the voluntariness criterion needs to separate true confessions from false confessions to the extent feasible. The second problem is law-enforcers’ position as self-seeking agents. There is no alignment between society’s interest in the conviction of the guilty and the exoneration of the innocent and the law-enforcers’ personal interests. The law-enforcers – police, prosecutors, and judges – have an incentive to boost their careers, prestige, and salaries while economizing their efforts. Confessions, guilty pleas, and asymmetrical information help them realize this goal. By eliciting a confession from a suspect, police officers can expedite the closing of the investigation. By obtaining a guilty plea from the defendant, the prosecutor can avoid an effort-consuming and unpredictable trial and successfully close the case. Confessions and guilty pleas also enable judges to streamline criminal proceedings and clear dockets.

Law-enforcers typically prefer true confessions and guilty pleas over false confessions and guilty pleas. This preference aligns with social good. But law-enforcers might also prefer false confessions and guilty pleas to p. 380effort-intensive and time-consuming investigations and trials. This preference does not align with social good. The law-enforcers may nonetheless pursue it. By doing so, they would impose serious agency costs on society.

Under asymmetrical information, no confession and guilty plea is demonstrably false or true. Any confession and guilty plea consequently can be claimed to be both true and false. The probabilities of those claims vary from case to case, with some claims being more persuasive than others. Yet, the vast majority of those claims are easy to make but difficult to refute. Defendants have an incentive to claim that they are innocent even when they are guilty. Their self-serving stories about making a false confession under pressure therefore will always be questioned. By the same token, police and prosecutors have an incentive to claim – both rightly and wrongly – that defendants’ confessions are true.

More often than not, judges will act upon a similar self-legitimizing motivation. A career-driven judge does not openly acknowledge that she admits confessions into evidence and accepts guilty pleas in order to streamline the proceedings and lock criminals in.

The credibility contest between the defendants’ and the law-enforcers’ self-serving claims is far from being equal. Law-enforcers will virtually always have an upper hand in this contest. As an initial matter, society gives criminal defendants very little credibility relative to that of law-enforcers (who protect it from crime). Police and prosecutors also have enormous resources and a far greater ability than defendants to gather evidence. Furthermore, police and prosecutors can allocate their resources as they deem appropriate. Consequently, they can threaten to use those superior resources against any recalcitrant defendant who refuses to confess and plead guilty. This strategy of extracting confessions and guilty pleas will be analogous to predatory pricing that exploits inequality in firms’ access to capital markets (Posner 1999, at 1505 and n. 59).

Credibility contests between defendants, on the one hand, and police and prosecution, on the other hand, are settled primarily by courts. Courts, however, have a strong incentive to streamline the adjudicative process by relying on confessions and guilty pleas. Doing so will economize the judge’s effort and help her establish a popular reputation for being “tough on crime.” As already indicated, these incentives make courts biased in favor of the police and the prosecution. This systemic bias can easily influence the jury as well. The judge’s decision to admit the defendant’s confession into evidence has a potential for over-influencing the jurors’ verdict. An average juror will treat any confession cleared by the judge as “admissible” as creditworthy. The confession consequently becomes indicative of the defendant’s guilt and gives the juror a good reason to return a guilty verdict.

p. 381To tackle these problems, the voluntariness requirement for confessions encompasses two sets of rules. The first set aims at separating the guilty from the innocent. To this end, it renders illegitimate any threat and favor capable of eliciting false confessions and guilty pleas. Any such threat and favor renders the confession “involuntary” and inadmissible. The second set of rules motivates law-enforcers to act as faithful agents for society. To achieve this outcome, the rules lay down a checklist by which courts must monitor police and prosecutorial misconduct. This compulsory monitoring motivates prosecutors to monitor police interrogations – a motivation that weakens the police’s ability to count on the prosecutor’s help in covering up misconduct.

3.2 Separating between True and False Confessions

The law separates between true and false confessions by setting up rules of admissibility and sufficiency. The admissibility rules specify and ban methods of interrogation capable of eliciting false confessions from innocent suspects. Those unlawful methods include violence, torture, and threats thereof, otherwise degrading and inhuman treatments, and severe psychological pressures. An average suspect experiencing any of those interrogation methods becomes willing to confess to the crime in order to avoid the suffering. This suspect may be either guilty or innocent: a guilty suspect will confess to the crime he actually committed; an innocent suspect will tell the interrogators anything they want him to say. The suspect’s confession consequently becomes unreliable. Instead of separating the guilty from the innocent, it pools them together. Any such confession therefore is inadmissible and can never be used as evidence against the defendant.

To be admissible, a defendant’s confession to the police must be free of such coercive interrogation methods. Many states require the prosecution to prove this fact to the trial judge beyond a reasonable doubt. Other jurisdictions, including federal courts, are satisfied by a preponderance of the evidence – the minimal constitutional proof requirement.14 Under both regimes, the trial judge must conduct a special hearing to examine the propriety of the defendant’s interrogation by the police. The judge will proceed on the assumption that a rational suspect, guilty or innocent, does not confess to the crime absent overwhelming reasons for making a confession. Based on this assumption, the judge will try to ascertain the reasons that prompted the defendant to speak against his own interest.

p. 382Judges will normally apply the voluntariness standard in a way most favorable to the defendant. The main reason for that is the rule against double jeopardy. Under this rule, acquittals are not appealable but convictions are. This asymmetric system of appeals skews errors in the application of the voluntariness standard (and some other legal requirements) against the prosecution. The trial judge is well aware of the fact that an error in admitting an involuntary confession into evidence will likely lead to a reversal of the defendant’s conviction by the appellate court. Following such reversal, the defendant cannot face another trial for the same crime: the rule against double jeopardy will prevent it (jeopardy attaches to a trial when the jury is empanelled and sworn, or, in bench trials, after the first witness has taken oath). On the other hand, the judge’s erroneous suppression of the defendant’s confession brings about no prospect of reversal because the prosecution cannot appeal. The asymmetrical appeal system interacts with judges’ fear of reversal. As such, it creates a strong pro-defendant pressure on the judges’ determinations of “voluntariness.” Facing this one-sided pressure, both trial and appellate courts tend to decide borderline cases in the defendant’s favor and rule his confession involuntary and inadmissible (Stith 1990).

To convince the judge that the defendant’s interrogation was not coercive, the prosecution normally would have to explain the confession’s rationality. Typically, the prosecution would have to show that the defendant confessed to the crime after being confronted with inculpatory evidence that persuaded him that denying the accusations is pointless. The prosecution therefore would need to adduce inculpatory evidence besides the confession. Another reason for having such independent evidence is the formal corroboration requirement, which is well-nigh universal. This requirement does not allow factfinders to find the defendant guilty on the basis of his confession alone. Evidence other than the defendant’s own words must verify the confession. Absent such evidence, the defendant would be entitled to a directed acquittal.

Accordingly, the prosecution and the police have a strong incentive to search for evidence credibly separating the guilty from the innocent. Arguably, this incentive is not strong enough to induce the desired separation. Many scholars believe that police and prosecutors need to have a more robust incentive for conscientiously working to eliminate erroneous convictions (Garrett 2010). To this end, some scholars have proposed a sentence reduction for every defendant whose conviction rests primarily on his confession to the police (Fisher and Rosen-Zvi 2008).

3.3 p. 383Monitoring Law-Enforcers

The historic Miranda decision15 laid down the currently prevalent four warnings requirement and the exclusionary rule. Miranda requires that, at the outset of a suspect’s custodial interrogation, the police advise the suspect of his right to remain silent; of the prospect that any part of his statement will be used as evidence against him in a criminal trial; of his right to consult an attorney and to have an attorney present at his questioning; and, finally, of the right to be represented by an attorney at the government’s expense when the suspect cannot afford to hire his own attorney. The suspect may remain silent indefinitely, or until he has consulted with an attorney and secured the attorney’s presence at the interrogation. Alternatively, the suspect can make a knowing, intelligent, and voluntary waiver of the Miranda rights and speak to the police. The police’s failure to follow Miranda leads to an automatic suppression of the suspect’s confession. A confession obtained in violation of Miranda is deemed involuntary and is consequently inadmissible. The Miranda exclusionary rule, however, does not extend to physical evidence that the police obtain with the help of the suspect’s unwarned – but not physically coerced – confession.16

This exclusionary rule performs an important prophylactic role: the police’s need to align their interrogations with Miranda minimizes the suspect’s prospect of being coerced into making a confession. Suppression of a suspect’s confession deters police misconduct better than do criminal punishment, disciplinary sanctions, and tort remedies. These alternative sanctions do not efficaciously detect and deter police misconduct. All of them require a separate proceeding – criminal, civil, or disciplinary – which makes them more expensive to administer than Miranda’s exclusionary rule. To secure those sanctions’ application, the defendant needs to convince the court (or a disciplinary tribunal) that he was a victim of coercive interrogation. Relative to a Miranda violation, those allegations are difficult to prove. Moreover, the victim’s remedies under Miranda’s p. 384alternatives can only be compensation for the damage suffered and moral vindication, as opposed to an acquittal or non-prosecution. Those remedies will often fail to provide a sufficiently strong incentive for the victim and his attorney to press charges against the police. To fix this incentive and to threaten the police with a high expected penalty for misconduct, the law needs to heighten the victim’s compensation and intensify the criminal and disciplinary punishments for defaulting police officers. But if those sanctions become an effective deterrent, there would be no evidentiary gains relative to the Miranda regime because the police would not have the evidence that Miranda presently suppresses.17

4. Waiver of the Self-Incrimination Privilege

As a general rule, defendants and suspects can waive any of their rights under the privilege against self-incrimination. The waiver, however, must be voluntary and informed in order to be considered effective. The same requirement applies to a plea of guilty or nolo contendere before the court. Before accepting any such plea, the court ought to make sure that the defendant fully understands the plea’s nature, is informed of his rights and is not acting under compulsion.18

Moreover, a defendant can waive his self-incrimination privilege together with the right (under Federal Rule of Evidence 410 and its state equivalents) to suppress any statement he made during plea negotiations and proffer sessions.19 Under the Mezzanatto rule, this waiver is effective (provided, again, that the defendant made it knowingly and voluntarily). Based on this waiver, the prosecution can use the defendant’s admissions to discredit his testimony and other evidence contradicting those admissions.

The Mezzanatto rule has a sound economic justification. In negotiating a plea bargain, the prosecution often needs to rely on the defendant’s representations, for example, on those that describe his part in the crime relative to other participants. The defendant’s entitlement to suppress anything he says makes his representations “cheap talk” upon which the prosecution cannot rationally rely. The prosecution consequently cannot engage in a meaningful negotiation with the defendant. Many defendants, however, find it in their interest to participate in proffer sessions and negotiate pleas with the prosecution. Society, too, has an obvious interest in substituting costly trials by plea bargains that streamline convictions. p. 385Defendants therefore need to be able to communicate with the prosecution credibly. To have credibility, they need to be able to make representations qualifying as “costly signals”: in order to be believed, the defendant must commit himself to a real and painful penalty for cheating. This is what the Mezzanatto rule does: it facilitates deals between criminal defendants and prosecution (Rasmusen 1998).

Suppression of such confessions is dictated not only by the Fifth Amendment’s privilege against self-incrimination, but also by common law, as well as by the defendant’s constitutional entitlement to due process. See Withrow v. Williams, 507 US 680, 693 (1993).

See Miranda v. Arizona, 384 US 436 (1966).

Both false positives and false negatives dilute deterrence by lowering the expected penalty for potential offenders. False negatives do so by reducing the offender’s probability of conviction. False positives do so by eroding the difference between the penalties expected from violating and not violating the law. See Polinsky and Shavell (2000), 60–62.

According to a well-known empirical study, Cassell and Fowles (1998), the right to silence reduces the conviction rate. This study examined the FBI’s case-clearance data that correlate with suspects’ confessions. Based on these data, the study estimated regression models for a variety of crime categories, using the clearance rate as the dependent variable. To capture the right’s effect on the clearance rate, the study created a dummy variable which equals 0 before 1966 (the year in which the Supreme Court decided Miranda v. Arizona, 384 US 436 (1966)) and 1 thereafter. This dummy variable was significant at the 0.01 level for serious crimes and was negatively signed, which means that the right to silence had significantly reduced the clearance rate. See id., at 1082–4. For methodological reservations about this study, see Donohue (1998), 1152–6 (casting doubts on whether clearance rates are dependable and methodologically adequate data for measuring the effects of Miranda on law enforcement). Also: federal law recognizes the right to silence and a silent defendant’s privilege against adverse inferences since 1943; see Johnson v. United States, 318 US 189, 198–9 (1943) (holding that, independent of the Fifth Amendment, a prosecutor cannot comment on the defendant’s invocation of the right to silence, and asserting that the Supreme Court’s supervisory power under Article III, § 1 of the United States Constitution makes it mandatory for federal courts to follow this holding), if not before; see Twining v. New Jersey, 211 US 78 (1908) (holding that the right to silence and a silent defendant’s privilege against adverse inferences belong to federal law, but are neither “privileges or immunities” nor “due process” within the meaning of the Fourteenth Amendment’s restrictions upon states). The Miranda Justices expressly modeled their warning requirements on the already established FBI practice of warning suspects: see Miranda, 384 US at 483–4 (attesting that “Over the years the Federal Bureau of Investigation has compiled an exemplary record of effective law enforcement while advising any suspect or arrested person, at the outset of an interview, that he is not required to make a statement, that any statement may be used against him in court, that the individual may obtain the services of an attorney of his own choice and, more recently, that he has a right to free counsel if he is unable to pay. A letter received from the Solicitor General in response to a question from the Bench makes it clear that the present pattern of warnings and respect for the rights of the individual followed as a practice by the FBI is consistent with the procedure which we delineate today.” (emphasis added)). As early as in 1952, the FBI’s director, J. Edgar Hoover, made an unequivocal statement (quoted in Miranda, 384 US at 483, n. 54) that “Special Agents are taught that any suspect or arrested person, at the outset of an interview, must be advised that he is not required to make a statement and that any statement given can be used against him in court. Moreover, the individual must be informed that, if he desires, he may obtain the services of an attorney of his own choice.” See Hoover (1952), 182. The 1966 baseline, chosen by Cassell and Fowles, therefore cannot be suitable for analyzing Miranda’s effects on the law-enforcement agencies in the states that followed the federal rules (and, of course, on the FBI).

See Bierschbach and Stein (2005), 1775–6.

The Supreme Court, however, has refused to recognize that the right to silence protects defendants against compelled production of handwriting samples. See Gilbert v. California, 388 US 263, 266–7 (1967). This refusal is the only part of the Fifth Amendment jurisprudence that does not align with the anti-pooling rationale: see Seidmann and Stein (2000), 477.

See United States v. Hubbell, 530 US 27 (2000).

See Stein (2005), 143–8.

An international treaty setting up cooperative law-enforcement between the United States and the foreign country might alter this analysis. See Seidmann and Stein (2000), at 483.

See New York v. Quarles, 467 US 649 (1984).

Adverse inferences, presumptions, and similar evidentiary devices are therefore no substitute for an aggressive discovery regime: see Stein (1996), 337–8.

For a good, but fictional, counterexample, see Fyodor N. Dostoyevsky, Crime and Punishment (1866) (Wordsworth, 2000).

The legal system can eliminate those fears and favors by making all confessions inadmissible as evidence and by abolishing guilty pleas. This extreme measure will dramatically increase the cost of criminal law-enforcement and intensify the pooling of guilty and innocent defendants.

See Lego v. Twomey, 404 US 477 (1972).

Miranda v. Arizona, 384 US 436 (1966).

United States v. Patane, 542 US 630 (2004). The Miranda exclusionary rule has another important limitation: the impeachment exception. Under this exception, if the defendant testifies contrary to his Miranda-barred confession, the prosecution can use that confession to impeach the defendant’s testimony in court: see Harris v. New York, 401 US 222 (1971). Note that this exception does not extend to unreliable confessions elicited by coercive interrogation methods: see Mincey v. Arizona, 437 US 385 (1978). The impeachment exception has a solid economic explanation: it curbs the defendants’ perverse motivation to use the exclusionary rule as a shield for perjury.

See Posner (1999), 1533.

See Federal Rule of Criminal Procedure 11(b).

United States v. Mezzanatto, 513 US 196 (1995).