discrimination is barred but even where it exists proof of discriminatory intent is impossible. 4 Laws need not be regularly enforced. Prosecutors can charge a handful of defendants and ignore hundreds of thousands of violators. The same proposition appears to hold for sentencing rules. Which leads to the third example of a constitutional dog not barking: the(non-regulation of plea bargaining. The Supreme Court has consistently refused to place even modest limits on the threats the government may use to extract guilty pleas. And thus far, the Court has declined to require that the government disclose exculpatory evidence prior to entry of a guilty plea. In criminal trials, the Constitution is omnipresent. In guilty pleas, it is nearly Invisible racially discriminatory private schools). Yet reading those cases, one senses that it is criminal prosecutors whose discretion the Court most wishes to protect 6 The key decisions are United States v. Armstrong, 517 U.S. 456(1996), and McCleskey v. Kemp, 481 U.S. 279(1987) See Wayte v. United States, 470 U.s. 598(1985)(16 out of 674,000 draft onregistrants prosecuted; Court held that prosecution did not violate due process or equal protection) Thus, the Court in Bordenkircher v Hayes, 434 U.S. 357(1978), did not inquire into the frequency(or not)of Kentucky prosecutors'enforcement of the state's three-strikes law, even though that laws application to Paul Hayes seemed strange at best, outrageous at worst. See infra notes 278-280 and accompanying text In Brady v United States, 397US 742, 743-44(1970), the defendant pled guilty in response to a threat to seek the death penalty. The statute authorizing the death penalty was later invalidated; the Supreme Court nevertheless held that the defendants plea was oluntary. Id at 756-58. The prosecutor in Bordenkircheroffered the defendant a plea with a five-year sentence; if the defendant refused, the prosecutor would seek a life sentence under a state three-strikes law. See 434 U.s. at 358-59. The defendant did refuse. the prosecutor carried out his threat, and the Supreme Court found the threat permissible. ld. The case most nearly on point is United States v Ruiz, 536 U.S. 622(2002), which rejected the defendants claim but left open the possibility that other claimants might fare better
racially discriminatory private schools). Yet reading those cases, one senses that it is criminal prosecutors whose discretion the Court most wishes to protect. 64The key decisions are United States v. Armstrong, 517 U.S. 456 (1996), and McCleskey v. Kemp, 481 U.S. 279 (1987). 65See Wayte v. United States, 470 U.S. 598 (1985) (16 out of 674,000 draft nonregistrants prosecuted; Court held that prosecution did not violate due process or equal protection). 66Thus, the Court in Bordenkircher v. Hayes, 434 U.S. 357 (1978), did not inquire into the frequency (or not) of Kentucky prosecutors’ enforcement of the state’s three-strikes law, even though that law’s application to Paul Hayes seemed strange at best, outrageous at worst. See infra notes 278-280 and accompanying text. 67In Brady v. United States, 397 U.S. 742, 743-44 (1970), the defendant pled guilty in response to a threat to seek the death penalty. The statute authorizing the death penalty was later invalidated; the Supreme Court nevertheless held that the defendant’s plea was voluntary. Id. at 756-58. The prosecutor in Bordenkircher offered the defendant a plea with a five-year sentence; if the defendant refused, the prosecutor would seek a life sentence under a state three-strikes law. See 434 U.S. at 358-59. The defendant did refuse, the prosecutor carried out his threat, and the Supreme Court found the threat permissible. Id. at 364-65. 68The case most nearly on point is United States v. Ruiz, 536 U.S. 622 (2002), which rejected the defendant’s claim but left open the possibility that other claimants might fare better. 15 discrimination is barred but even where it exists, proof of discriminatory intent is impossible.64 Laws need not be regularly enforced. Prosecutors can charge a handful of defendants and ignore hundreds of thousands of violators.65 The same proposition appears to hold for sentencing rules.66 Which leads to the third example of a constitutional dog not barking: the (non-)regulation of plea bargaining. The Supreme Court has consistently refused to place even modest limits on the threats the government may use to extract guilty pleas.67 And thus far, the Court has declined to require that the government disclose exculpatory evidence prior to entry of a guilty plea.68 In criminal trials, the Constitution is omnipresent. In guilty pleas, it is nearly invisible
B. Legislative politics and the Constitution When the Supreme Court constitutionalized criminal procedure in the 1960s the conventional wisdom, evidently shared by the Justices, held that elected legislators would never adequately protect the interests of criminal suspects and defendants. That view seemed plausible at the time-though the contrary view was also plausible, as Corinna Lain and Orin Kerr have shown. Today, the Justices political prophecy looks either wrong or self-fulfilling. Members of Congress and state legislators can and do protect the interests of the criminal justice systems targets. That protection follows some basic patterns. Legislators are more likely to craft fair procedures than to write fair substantive laws, and more likely to guard against abusive policing than to prevent excessive punishment. The Court drove legislators, along with the dollars they control, away from those areas where legislation might have done the most good(policing and procedure), and into those areas where it is bound to do the most harm(crime definition and sentencing) Three propositions are key. First, constitutional law makes legislative regulation of constitutionalized subjects politically costly. Second, for the past few decades, constitutional law has made legislation in constitutionally unregulated areas politically cheap. Third, constitutional law distorts state and federal spending Congress and state legislatures spend where they can also govern, and constitutional law makes governing policing and trial procedure difficult 1. Policing and Procedure Constitutional rules of policing and trial procedure are mandatory; legislators may add to but cannot replace them. At a glance, that seems obviously right -of course legislators cannot substitute their own preferred rules for constitutional mandates- but it is far from inevitable, and as a matter of political economy it is probably perverse. Default rules would encourage legislative protection of For the general argument, see Corinna Barrett Lain, Countermajoritarian Hero or Zero? Rethinking the Warren Courts Role in the Criminal Procedure Revolution, 152 PA L. REV. 1361(2004). Kerr's article nicely complements Lain's: Kerr shows that legislative privacy protection has often worked better than the judicially mandated kind along the way he offers a telling history of Congressional regulation of wiretapping, both before and during the Warren era. See Orin S. Kerr, The Fourth Amendment and New Technologies: Constitutional Myths and the Case for Caution, 102 MICH. L REv. 801, 839 50(2004)[hereinafter Kerr, Constitutional Myths
69For the general argument, seeCorinna Barrett Lain, Countermajoritarian Hero or Zero? Rethinking the Warren Court’s Role in the Criminal Procedure Revolution, 152 U. PA. L. REV. 1361 (2004). Kerr’s article nicely complements Lain’s: Kerr shows that legislative privacy protection has often worked better than the judicially mandated kind; along the way he offers a telling history of Congressional regulation of wiretapping, both before and during the Warren era. See Orin S. Kerr, The Fourth Amendment and New Technologies: Constitutional Myths and the Case for Caution, 102 MICH.L.REV. 801, 839- 50 (2004) [hereinafter Kerr, Constitutional Myths]. 16 B. Legislative Politics and the Constitution When the Supreme Court constitutionalized criminal procedure in the 1960s the conventional wisdom, evidently shared by the Justices, held that elected legislators would never adequately protect the interests of criminal suspects and defendants. That view seemed plausible at the time — though the contrary view was also plausible, as Corinna Lain and Orin Kerr have shown.69 Today, the Justices’ political prophecy looks either wrong or self-fulfilling. Members of Congress and state legislators can and do protect the interests of the criminal justice system’s targets. That protection follows some basic patterns. Legislators are more likely to craft fair procedures than to write fair substantive laws, and more likely to guard against abusive policing than to prevent excessive punishment. The Court drove legislators, along with the dollars they control, away from those areas where legislation might have done the most good (policing and procedure), and into those areas where it is bound to do the most harm (crime definition and sentencing). Three propositions are key. First, constitutional law makes legislative regulation of constitutionalized subjects politically costly. Second, for the past few decades, constitutional law has made legislation in constitutionally unregulated areas politically cheap. Third, constitutional law distorts state and federal spending. Congress and state legislatures spend where they can also govern, and constitutional law makes governing policing and trial procedure difficult. 1. Policing and Procedure Constitutional rules of policing and trial procedure are mandatory; legislators may add to but cannot replace them. At a glance, that seems obviously right — of course legislators cannot substitute their own preferred rules for constitutional mandates — but it is far from inevitable, and as a matter of political economy it is probably perverse. Default rules would encourage legislative protection of
constitutional interests. If Miranda were a default rather than a mandatory rule, 70 legislators who dislike Miranda might join with legislators who want better limits on police coercion to pass, say, laws requiring video-and audiotaping of interrogation sessions. That is how constitutional defaults work: they expand support for alternative regulatory strategies Mandatory rules have the opposite effect. The existence of any given constitutional rule tends to reduce support for friendly legislation. Like the marginal benefit of most things, the marginal benefit of regulation declines; each new increment is worth less than the one before. One can imagine a number of different legal rules aimed at stopping coercive police interrogation: Miranda-style warnings, mandatory taping, a list of forbidden interrogation tactics, and no doubt there are others. Any of those rules would likely stop the most egregious police misconduct Once one rule is in place, therefore, the benefit of adding another is bound to shrink That pattern holds throughout criminal procedu As marginal benefit falls, marginal cost rises. The government pays for criminal procedure rules in the coin of foregone arrests and convictions. When a particular rule turns winning cases into losers, prosecutors seek substitutes: other cases to take the place of the ones lost. Naturally, the highest-value substitutes 7 Earl Warren hinted that it might be a default. See Miranda v. Arizona, 384 U. S 436, 467(1966)(inviting legislators to"exercise.. their creative rule-making capacities") But the hint was never serious as the court confirmed in dickerson v United States. 530 U.S. 428(2000)(invalidating a congressional effort to overrule Miranda) Mandatory taping probably would be the equilibrium regulatory strategy wer Miranda a default rule rather than a mandatory rule. Illinois, Maine, Texas, and the District of Columbia require the taping of interrogation sessions by statute: Alaska and Minnesota have a similar requirement via judicial decision. More than a dozen other states have considered mandatory taping bills in the past two years. See Joelle Anne Moreno, Faith- Based Miranda? Why the New Missouri v Seibert"Bad Faith"Test is a Terrible ldea, 47 ARIZ. L. REv. 395, 418& nn. 176-201(2005). It seems reasonable to suppose that, were Miranda's restrictions relaxed, a great many more such laws would have been enacted For a similar argument in a different context, see Einer Elhauge, Preference Eliciting Statutory Default Rules, 102 COLUM. L REV. 2162(2002) For a discussion of how crime rates and resource constraints affect the choice of substitutes, see William J Stuntz, The Uneasy Relationship Between Criminal Procedure
70Earl Warren hinted that it might be a default. See Miranda v. Arizona, 384 U.S. 436,467 (1966) (inviting legislators to “exercise . . . their creative rule-making capacities”). But the hint was never serious, as the Court confirmed in Dickerson v. United States, 530 U.S. 428 (2000) (invalidating a congressional effort to overrule Miranda). 71Mandatory taping probably would be the equilibrium regulatory strategy were Miranda a default rule rather than a mandatory rule. Illinois, Maine, Texas, and the District of Columbia require the taping of interrogation sessions by statute; Alaska and Minnesota have a similar requirement via judicial decision. More than a dozen other states have considered mandatory taping bills in the past two years. See Joelle Anne Moreno, FaithBased Miranda?: Why the New Missouri v. Seibert “Bad Faith” Test is a Terrible Idea, 47 ARIZ. L. REV. 395, 418 & nn. 176-201 (2005). It seems reasonable to suppose that, were Miranda’s restrictions relaxed, a great many more such laws would have been enacted. 72For a similar argument in a different context, see Einer Elhauge, PreferenceEliciting Statutory Default Rules, 102 COLUM. L. REV. 2162 (2002). 73For a discussion of how crime rates and resource constraints affect the choice of substitutes, see William J. Stuntz, The Uneasy Relationship Between Criminal Procedure 17 constitutional interests. If Miranda were a default rather than a mandatory rule,70 legislators who dislike Miranda might join with legislators who want better limits on police coercion to pass, say, laws requiring video- and audiotaping of interrogation sessions.71 That is how constitutional defaults work: they expand support for alternative regulatory strategies.72 Mandatory rules have the opposite effect. The existence of any given constitutional rule tends to reduce support for friendly legislation. Like the marginal benefit of most things, the marginal benefit of regulation declines; each new increment is worth less than the one before. One can imagine a number of different legal rules aimed at stopping coercive police interrogation: Miranda-style warnings, mandatory taping, a list of forbidden interrogation tactics, and no doubt there are others. Any of those rules would likely stop the most egregious police misconduct. Once one rule is in place, therefore, the benefit of adding another is bound to shrink. That pattern holds throughout criminal procedure. As marginal benefit falls, marginal cost rises. The government pays for criminal procedure rules in the coin of foregone arrests and convictions. When a particular rule turns winning cases into losers, prosecutors seek substitutes: other cases to take the place of the ones lost.73 Naturally, the highest-value substitutes go
first. As regulation piles up and more cases fall, the substitutes are worth less and less. The cost of the lost cases grows. At some point, the substitutes run out and the cost jumps sharply- lost cases are simply lost; there are no replacements. So the marginal cost curve looks like a partially flattened S [Figure 1] The more constitutional law regulates, the more likely legislators are to find selves in In theory, the costs of different regulatory options might overlap if the overlap were large enough, the marginal cost of adding new regulation would be small. The reality is different. Mirandas costs are felt when suspects invoke their rights immediately after hearing the famous warnings. Mandatory taping would lead to lost confessions in cases involving long, drawn-out questioning of vulnerable suspects. A list of banned tactics would cost the police confessions in cases in which those tactics would be especially useful, or perhaps especially tempting to the police These sets of lost cases overlap only slightly Take another example. One might regulate street stops by mandating individualized suspicion, as current Fourth Amendment law does, or by an antidiscrimination rule. Requiring individualized suspicion encourages the police to shift resources from middle-class neighborhoods, where drug crime takes place and Criminal Justice, 107 YALE L J. 1, 23-27(1997)[hereinafter Stuntz, Uneasy Relationship Roughly 20% of interrogated suspects invoke their Miranda rights, thus barrin further questioning. Only 1% of suspects invoke their rights after questioning has begun the rest of the invocations come when the Miranda warnings are first read to the suspect. For the leading study, see Richard A Leo, The Impact of Miranda Revisited, J. CRIM. L.& CRIMINOLOGY 621, 653(1996)
and Criminal Justice, 107 YALE L.J. 1, 23-27 (1997) [hereinafter Stuntz, Uneasy Relationship]. 74Roughly 20% of interrogated suspects invoke their Miranda rights, thus barring further questioning. Only 1% of suspects invoke their rights after questioning has begun — the rest of the invocations come when the Miranda warnings are first read to the suspect. For the leading study, see Richard A. Leo, The Impact of Miranda Revisited, J. CRIM. L. & CRIMINOLOGY 621, 653 (1996). 18 first. As regulation piles up and more cases fall, the substitutes are worth less and less. The cost of the lost cases grows. At some point, the substitutes run out and the cost jumps sharply — lost cases are simply lost; there are no replacements. So the marginal cost curve looks like a partially flattened S: [Figure 1] The more constitutional law regulates, the more likely legislators are to find themselves in the expensive portion of that curve. In theory, the costs of different regulatory options might overlap; if the overlap were large enough, the marginal cost of adding new regulation would be small. The reality is different. Miranda’s costs are felt when suspects invoke their rights immediately after hearing the famous warnings.74 Mandatory taping would lead to lost confessions in cases involving long, drawn-out questioning of vulnerable suspects. A list of banned tactics would cost the police confessions in cases in which those tactics would be especially useful, or perhaps especially tempting to the police. These sets of lost cases overlap only slightly. Take another example. One might regulate street stops by mandating individualized suspicion, as current Fourth Amendment law does, or by an antidiscrimination rule. Requiring individualized suspicion encourages the police to shift resources from middle-class neighborhoods, where drug crime takes place
indoors, to poor ones, where open-air drug markets and street gangs operate. (The government can moreeasily show the requisite suspicion when officers witness some portion of the crime. Because of the strong correlation between race and urban poverty, a serious antidiscrimination rule would cut in the opposite direction making street stops cheaper in wealthy areas and more expensive in poorer ones These two forms of regulation have no substantial cost overlal Or, consider trial procedure. a broad jury trial right costs the government convictions of defendants who are most appealing to jurors. Raising the burden of proof costs cases in which the government can meet a lower burden but not the igher one. Both rules seek greater accuracy, but they seek that goal in different ways, so their cost is felt in different cases. If constitutional law adopts one of these rules, the marginal cost of the second is higher than it would have been given a constitutional blank slate The bottom line is simple. Constitutional regulation raises the political price of legislative regulation. But how much does the higher price matter? Perhaps the answer is, not much. Legislative regulation of policing and trial procedure might be politically impossible regardless of what the Supreme Court does. If so, the political costs of criminal procedure doctrines are beside the point Both theory and evidence suggest otherwise. Legislators respond to powerful interest groups. Contrary to the conventional wisdom, criminal suspects are a powerful interest group. The police stop 23 million motorists per year. No one knows how many pedestrians the police stop but that number is probably higher CoLlAFor a more extended argument, see William J Stuntz, Race, Class, and Drugs, 98 L. REV. 1795, 1804-15(1998)[hereinafter Stuntz, Race and Drugs. 7For the classic treatment. see WILLIAM JULIUS WILSON THE TRULY DISADVANTAGED(1987). For more recent discussions, see Robert J. Sampson William Julius Wilson, Toward a Theory of Race, Crime, and Urban Inequality, in CRIME AND INEQUALITY 41(John Hagan Ruth D. Peterson eds, 1995); Michael H Schill& Susan M Wachter, The Spatial Bias of Federal Housing Law and Policy: Concentrated Poverty in Urban America, 143 U PA L REV. 1285(1995) 72002 STATISTICAL ABSTRACT, supra note 4, at 195 No. 309
75For a more extended argument, see William J. Stuntz,Race, Class, and Drugs, 98 COLUM. L. REV. 1795, 1804-15 (1998) [hereinafter Stuntz, Race and Drugs.] 76For the classic treatment, see WILLIAM JULIUS WILSON, THE TRULY DISADVANTAGED (1987). For more recent discussions, see Robert J. Sampson & William Julius Wilson, Toward a Theory of Race, Crime, and Urban Inequality, in CRIME AND INEQUALITY 41 (John Hagan & Ruth D. Peterson eds., 1995); Michael H. Schill & Susan M. Wachter, The Spatial Bias of Federal Housing Law and Policy: Concentrated Poverty in Urban America, 143 U. PA. L. REV. 1285 (1995). 772002 STATISTICAL ABSTRACT, supra note 4, at 195 No. 309. 19 indoors, to poor ones, where open-air drug markets and street gangs operate.75 (The government can more easily show the requisite suspicion when officers witness some portion of the crime.) Because of the strong correlation between race and urban poverty, 76 a serious antidiscrimination rule would cut in the opposite direction, making street stops cheaper in wealthy areas and more expensive in poorer ones. These two forms of regulation have no substantial cost overlap. Or, consider trial procedure. A broad jury trial right costs the government convictions of defendants who are most appealing to jurors. Raising the burden of proof costs cases in which the government can meet a lower burden but not the higher one. Both rules seek greater accuracy, but they seek that goal in different ways, so their cost is felt in different cases. If constitutional law adopts one of these rules, the marginal cost of the second is higher than it would have been given a constitutional blank slate. The bottom line is simple. Constitutional regulation raises the political price of legislative regulation. But how much does the higher price matter? Perhaps the answer is, not much. Legislative regulation of policing and trial procedure might be politically impossible regardless of what the Supreme Court does. If so, the political costs of criminal procedure doctrines are beside the point. Both theory and evidence suggest otherwise. Legislators respond to powerful interest groups. Contrary to the conventional wisdom, criminal suspects are a powerful interest group. The police stop 23 million motorists per year.77 No one knows how many pedestrians the police stop, but that number is probably higher