Full opinion text
MEMORANDUM, ORDER, AND JUDGMENT WEINSTEIN, Senior District Court J. Table of Contents: I. Introduction.417 II. Facts.418 A. Background.418 B. Procedural History.419 III. Law.422 A. Conflicts of Interest.422 B. Plea Bargains.422 1. Plea Bargaining and Coercion.422 2. Competency to Plead.426 3. Changes in the Criminal Justice System .427 C. Bail.434 1. Law Governing Bail.434 2. History of Bail Reform .434 3. Constitutional Requirements .437 4. Use of Pretrial Services.438 D. Speedy Trial.438 1. Constitutional Aspects.438 2. Statutes and Rules.439 E. Sentencing.439 1. Departure.439 a. Destruction of Livelihood .440 b. Stress of Pretrial Incarceration.441 2. Fine.442 3. Credit for Good Behavior.443 4. Supervised Release.443 IV. Application of Law to Facts.444 A. Defendant’s Ability to Plead .444 B. Conflicts of Interest.444 C. Sentencing.445 1. Calculation of offense level.445 2. Departure.445 3. Credit for time served and for good behavior.446 4. Supervised Release.446 V. Conclusion.446 Appendix 1: Types of Sentencing Departure by Year.446 Appendix 2: Disposition of Federal Criminal Defendants, National.448 Appendix 3: Disposition of Federal Criminal Defendants, Eastern District of New York.450 I. Introduction The questions posed are whether acceptance of a plea of guilty and a departure downward in sentencing were appropriate in this substantial money laundering case. Serious factual and legal issues are presented by evidence of the defendant’s long pretrial incarceration under onerous conditions, the government’s control over funding for her counsel and the threat of much greater punishment without a plea. This case illustrates the danger of due process violations by intensive pressure on defendants to plead guilty because of lengthy pretrial incarcerations and the offer of advantageous deals for lesser terms of imprisonment. The stick and carrot— largely controlled by prosecutors — produces a danger of excessive coercion of a defendant and undue pressures on defense counsel to avoid trial. It requires particularly close supervision by the court to ensure voluntariness of the plea. In evaluating the need for sentencing departures the unusual tensions some defendants face while in custody awaiting trial also necessitate consideration. The virtual elimination of federal criminal trials, substituting administrative decisions not to prosecute or pleas of guilty, has substantially changed our federal criminal law system. Increased prosecuto-rial discretion and power have raised the percent of guilty pleas from 86% of all federal convictions in 1971 to 95% in 2001. Discretion not to prosecute is widely exercised. Enhancement of control of sentencing by the prosecutor as a result of sentencing guidelines and minimum sentences has increased the government’s power to coerce defendants. There has been a change from the paradigmatic concept of investigation and accusation by the government of almost all persons believed to have committed crimes, trial by jury with a strong role for defense counsel, and discretion in sentencing by the court, to a system sharply reducing the role of defense counsel, the jury and the judge, and whatever protections they can afford a defendant. Defendant Hebroni and codefendants were charged with using jewelry businesses in Panama to launder some $10,000,000 of drug money. The prosecution was controlled by the Narcotics and Dangerous Drugs section of the United States Department of Justice, from Washington, D.C., rather than by the district’s United States Attorney. After lengthy pre-trial proceedings, drawn out over a period of more than a year and a half largely because of failure of the government to produce the huge amount of documents seized by the government of Panama; successive grants of bail by the trial court which were overturned on appeal; and strong government opposition to payment of defense counsel from defendant’s known assets, all of which had been seized; defendant and the government entered into a plea agreement. It required defendant to be incarcerated for a term of somewhere between 38 and 41 months under the guidelines, and to give up all her assets. Without such an agreement, upon conviction she would have faced a prison sentence of 151 to 188 months and possibly more, based on a level of 34 or higher. See U.S. Sentencing Guidelines Manual §§ 2Sl.l(b)(l) & 2S1.1 (b)(2)(I) (2001) (hereinafter “Guidelines Manual”). The court entertained doubts about 1) defendant’s capacity and desire to sign the plea agreement and 2) a possible conflict of interest of defense counsel in recommending a plea. Accordingly, the defendant was ordered to undergo a psychiatric evaluation to determine her fitness. Additional defense counsel was appointed to ensure that defendant’s rights were fully protected and that she was being appropriately advised. Ultimately defendant’s plea was accepted by the court. While defendant agreed not to seek a downward departure, the agreement did not limit the court’s independent authority to consider such a departure. On the basis of the presentence report, briefs by both sides and sentencing hearings, the court departed downward 2 levels and sentenced the defendant to a total of 27 months in prison and forfeiture of all her personal and business assets, a fine, and a term of supervised release of three years, during which time defendant may not engage in the jewelry or rare metals business, nor conduct any business in Panama. II. Facts A Background Defendant Hebroni is a citizen of Israel domiciled in Panama. She is forty-nine years old and recently widowed. Her immediate family is her 6-year-old son, now a resident of Israel, who was conceived after many years of medical intervention. Defendant appears to be an intelligent woman, though physically frail. The court was presented with no admissible evidence. From the contentions of government counsel and concessions by defendant, the facts can be briefly summarized as follows: Co-defendant Speed Joyeros, S.A., is a jewelry business fully owned by defendant Hebroni. Speed Joyeros, S.A., engaged in the sale of gold and silver jewelry and precious metals to many retail and wholesale customers throughout Central and South America and in Europe and the Middle East. Defendant’s business was among the most successful jewelry businesses in the Colon, Panama Free Zone, with revenues of $155 million in 1998 and $105 million in 1999. She received cash, checks and electronic money transfers in a money laundering scheme. A primary modus was to sell jewelry to drug lords, knowing that it was being paid for with drug money, thus allowing them to convert dirty money into glistening clean jewelry. Financial institutions in the United States and other countries were employed in these conversions. Co-defendant Argento Vivo S.A. is a corporation controlled by defendant’s brother and a named co-defendant, Eliahu Mizrahi, who has not been found. Argento Vivo, S.A. is a wholesale and retail jewelry dealer engaged in the sale of silver jewelry to retail stores located throughout Central and South America. B. Procedural History On September 22, 2000, Hebroni was arrested and detained on money laundering charges. She had voluntarily come to this country from Panama to defend against forfeiture proceedings commenced here against millions of dollars of her assets, mainly in Panama. Some $ 900,000 were in American banks. The indictment charged the defendant with five counts of washing and conspiring to wash drug money. 18 U.S.C. §§ 1956, 1957. A superced-ing indictment was filed more than a year later, on December 13, 2001, clarifying the government’s legal theory. On October 10, 2000, defendant pled not guilty. After various pretrial motions, defendant sought bail. On July 2, 2001, bail was denied by the judge previously assigned to the case. The case lagged, with large numbers of the basic business records from Panama unavailable in this country. All the documents used by defendants in their Panamanian businesses had been seized by the government of Panama. Despite repeated attempts by the defendant to retrieve and examine these many thousands of papers in preparation for trial, they had still not been fully examined by the defense up to the time of plea. After new bail hearings, on December 12, 2001, defendant was ordered released if she could satisfy specified conditions. The magistrate judge’s hearings determined that pledged property fit the conditions set by the court. It was the court’s view that the bail and other conditions would ensure that defendant could prepare for trial and that she would be present at trial. These reasons were stated orally on the record. Transcript of Dec. 12, 2001, pp. 39-56. The court was particularly concerned that without bailed release there was a serious possibility that defendant would not be able to properly defend herself. Her defense as well as the prosecution was to be based on the documents seized from her place of business by Panamanian authorities. They had not yet been provided to her. In addition, her incarceration limited her ability to work with her attorneys, accountants, and the necessary calculators and computers required to analyze the extensive business records in this complex case. Trial was set for March 11, 2002, giving defendant just under three months to prepare. The government appealed from the order of December 12, 2001 granting release on conditions. Release was stayed by the court of appeals. By its decision of January 8, 2002, it vacated the district court’s order. The district court was directed to consider “the determinative question specified by § 3142 — whether any combination of conditions would reasonably assure He-broni’s appearance as required.” See United States v. Hebroni 25 Fed.Appx. 85, 86 (2d Cir.2002). Following remand, the district court held two additional evidentia-ry hearings. Important events bearing on defendant’s ability to defend had occurred. First, over strong government objection, the court ordered the release of sufficient of the defendants’ funds to pay her attorneys for work done to that time. Second, the government began to make available to defendant numerous boxes of papers. Since she was unable to study these documents except while she was detained, the court made available a room in the courthouse. She was brought to and from prison each day to view her business writings under the guard of the marshals. This procedure was burdensome since defendant was given limited space and she did not have the necessary equipment and assistance. Trial would have required introduction and analysis of thousands of documents, tracing transactions running through institutions in many countries. Third, defendant suffered physically. She submitted evidence indicating that she was being held in the open in cold weather for lengthy periods each day with inadequate clothing awaiting transportation between jail and court, and was not being provided with appropriate food. The court observed her physical and emotional deterioration. On February 1, 2002, the court ordered that she be provided with adequate clothing. On February 11, 2002, the court ordered that she be provided with appropriate food. Despite these measures, continued suffering of the defendant was noted. Conditions of incarceration of the defendant and the nature of the case raised doubts that due process could be served without some form of release. The defendant had to be able to prepare for her impending trial. In the court’s view the bail terms were sufficient to assure her presence at trial. The district court again granted bah on February 12, 2002 with detailed findings as required by the court of appeals. See United States v. Hebroni, 187 F.Supp.2d 75 (E.D.N.Y.2002). The conditions were stringent. Defendant was to be detained under house arrest; monitored by Pretrial Services with an electronic bracelet at her own cost; subject to telephone monitoring; subject to a curfew; guarded at all times if the government wished; required to place her passport in government custody; and required to post bail of $10,000,000 guaranteed by substantial real property. At the time bail was granted this second time, the court noted that defendant was only a month away from trial, so that further detention would seriously hinder her ability to present a defense. Delays in release on bail worked in the government’s favor since it shortened the time when defendant and her counsel could analyze the evidence in an appropriate environment. The government appealed from the district court’s February 12, 2002 order granting release on bail. The court of appeals stayed the order. Ultimately the court of appeals, on March 13, 2002, reversed the district court’s order “for the reasons stated in the government’s brief.” See United States v. Hebroni, 2002 WL 391179 (2d Cir. Slip Op. Mar. 13, 2002). As the defense’s difficulty in adequately preparing for trial under these conditions became more evident, and the date of trial grew closer, the parties notified the court that they wished to enter into a plea agreement. Defendant agreed to plead guilty to count one of the indictment. See 18 U.S.C. § 1956(h). Count one contains allegations of millions of dollars in money laundering, but the parties agreed that the guilty plea would be limited to $474,000 of specific money laundering events. Defendant agreed “not to contest that the Government would have proven beyond a reasonable doubt at trial in this case that the funds identified in Count One of the indictment were derived from drug trafficking.” Plea paragraph 5(b). The agreement incorporated contemplated terms of a sentence for the defendant. As already noted, the prospective guideline was at level 20, carrying a range of imprisonment of 33 to 41 months; it included a downward adjustment for acceptance of responsibility. Defendant agreed not to seek a downward departure or to appeal her sentence if it was not greater than 41 months. She also agreed to forfeit all of her own ownership of assets in the defendant companies (estimated at some $6,000,000) which had been seized by Panamanian authorities, as well as the remaining more than $600,0000 of assets in a United States bank. The indicted corporations also pleaded guilty as part of the deal. A hearing was held on March 1, 2002 to review the plea. For reasons set out orally at that time, and suggested in further detail in this memorandum, the court ordered that the defendant be examined by a physician to assist in determining if her plea was voluntary. The resulting report of Dr. Sanford L. Drob was illuminating. Dr. Drob noted that Ms. Hebroni had suffered physically during her incarceration, especially in recent months when she was shuttled back and forth from jail to court daily to review papers. He also pointed out that Ms. Hebroni was exceedingly anxious to be reunited with her infant son. She told Dr. Drob that the reason she had agreed to a plea was to be reunited with her son sooner. She was, according to the Doctor, under the erroneous impression that she would be sentenced to time served. She stated, according to his report, that she believed that “you can’t win” at trial. Dr. Drob concluded that defendant would have been far less likely to accept a plea agreement if she had been allowed to leave jail on bad while awaiting trial. Dr. Drob noted that “several factors” were involved in defendant’s decision to plead. The primary reason was concern for her son. She had suffered unusual stress as a result of separation from him— her only child and the result of years of fertility treatments. Since she was widowed, her son depended on her as his only parent; she had already had to explain an absence of 18 months. She also felt physically worn down and ill-equipped to undergo a lengthy trial. Since the report indicated that defendant had an erroneous understanding of the plea agreement — which did not, as she reportedly thought, require immediate release — the court ordered defense counsel to make clear to their client the effect of the agreement. The court appointed supplemental counsel to ensure that defendant’s rights were adequately protected. It was concerned that the lack of defendant’s assets to pay legal and other costs of a trial might have produced an inclination of defense counsel to reduce their own obligations and potential financial risk were there a full trial, by recommending a plea. On March 20, 2002, the court held a second hearing on the plea agreement. All parties, including supplemental counsel, were present. Defense counsel stated that defendant was aware of the effect of the plea agreement and was capable of making the agreement. Based on this affirmation; the medical report; an oral report of supplemental defense counsel that defendant understood the plea and that her counsel’s advice was based solely on her welfare; and close questioning of the defendant by the court, the defendant’s plea was accepted. See Transcript of Mar. 20, 2002 at 35-37. The government requested an adjournment of sentence to consider potential bases for a downward departure outlined by the court. Both sides submitted briefs on departure issues, the government contesting the court’s power to depart downward. A presentence report was presented by Probation. The report was significant in that it calculated defendant’s offense level differently than the plea agreement. The presentence report indicated that defendant’s adjusted offense level would be 24, rather than 20, requiring a minimum sentence of 51 months rather than the 33 months set out in the plea agreement. III. Law A. Conflicts of Interest Attorneys are subject to strict rules designed to prevent them from having conflicts of interest with their clients. The American Bar Association Model Code of Professional Responsibility demands that a lawyer “shall not accept [or continue in] employment if the exercise of his professional judgment on behalf of his client will be or reasonably may be affected by his own financial, business, property, or personal interests.” ABA Model Code DR 5-101. This principle is one of the pillars of our judicial system—that an attorney should not have any reservation in giving unalloyed support to the client he is representing. A lawyer constrained by a conflict of interest between attorney and client, or between two or more clients, may be required to withdraw as counsel. ABA Model Rule 1.16. The right to counsel free from conflicts of interest has been enforced by the Supreme Court. See Holloway v. Arkansas, 435 U.S. 475, 489-90, 98 S.Ct. 1173, 55 L.Ed.2d 426 (1978) (right to effective counsel includes right to attorney free from conflicts of interest). Forfeiture cases may be particularly amenable to conflicts of interest. In these cases, all of a defendant’s assets may be seized, preventing the attorney from recovering any fee if the defendant loses. The attorney has a financial incentive to ensure that fees are paid, and thus an incentive to advise a defendant in a manner likely to protect fees—a position which may diverge from the best interests of the defendant. All forfeiture cases do not raise conflicts per se. See, e.g., United States v. Monsanto, 491 U.S. 600, 615, 109 S.Ct. 2657, 105 L.Ed.2d 512 (1989) (finding no coercion in forfeiture case). Coercion must be measured by the way the government exercises its power to seize assets. See, e.g., United States v. Marquez, 909 F.2d 738 (2d Cir.1990) (“Any attempt to use the prospect of getting the defendant’s lawyer paid from seized funds as a bargaining chip to obtain a concession from the defendant poses the potential for a serious conflict of interest.”); cf. Cuyler v. Sullivan, 446 U.S. 335, 343-45, 100 S.Ct. 1708, 64 L.Ed.2d 333 (1980) (states may not conduct trials in a way that unconstitutionally impairs a defendant’s right to effective counsel). B. Plea Bargains 1. Plea Bargaining and Coercion The Federal Rules of Criminal Procedure allow a defendant to “plead guilty, not guilty, or nolo contendere.” F.R.Cr.P. 11(a)(1). Any plea of guilty must be shown to the court to be “voluntary”: The court shall not accept a plea of guilty or nolo contendere without first, by addressing the defendant personally in open court, determining that the plea is voluntary and not the result of force or threats or of promises apart from a plea agreement. The court shall also inquire as to whether the defendant’s willingness to plead guilty or nolo con-tendere results from prior discussions between the attorney for the government and the defendant or the defendant’s attorney. F.R.Cr.P. 11(d). While the rule warns of coercion as a “result of force or threats or of promises apart from a plea agreement,” the court’s obligation to determine “volun-tariness” is not limited by these examples. The word “voluntary” is itself laden with ambiguity; it is an antonym of “coerced” which may be due to various forms of non-force or non-threats. The second sentence starting with the words, “the court shall also” inquire about prior “discussions,” suggests that inappropriate coercion may exist even where there is no “force or threats or ... promises apart from a plea agreement.” There is no single clear definition of “voluntary” for all legal purposes. Even in the criminal-law-plea context, it is unclear whether “voluntary” means freedom from any coercion or whether it means freedom only from “wrongful” or “undue” coercion. A pristine rule of “no coercion” would preclude many plea agreements. Requiring plea negotiations to be free from “any coercion” would contradict the basic notions of bargaining. Contract law theory suggests that no bargaining process is completely devoid of coercion in some form. The problem is one of discerning what level of coercion is so inappropriate as to render a plea agreement invalid. The seventh edition of Black’s Law Dictionary’s definition of “voluntary” as something done “unconstrained by interference. . .not compelled by outside influence,” is not helpful in determining whether a guilty plea was “voluntary” since a defendant is always “influenced” by many factors including family demands and other social pressures as well as by the hope of minimizing punishment. To conclude that a plea agreement is made “unconstrained by interference” or “not compelled by outside influence” would be to ignore the reality that such an agreement is a bargain made between a relatively powerless defendant and a prosecutor who can exercise a great deal of influence over the accused’s future happiness. In Brady v. United States, 397 U.S. 742, 90 S.Ct. 1463, 25 L.Ed.2d 747 (1970), the Supreme Court explained that a plea agreement is based upon mutual advantages gained by both the prosecutor and the defendant. If the defendant is “fully aware of the direct consequences, including the actual value of any commitments made to him by the court, prosecutor, or his own counsel” the plea may be considered “voluntary.” Id. at 755, 90 S.Ct. 1463. Courts recognize that “a great many [defendants are] no doubt motivated at least in part by the hope or assurance of a lesser penalty than might be imposed if there were a guilty verdict after a trial to judge or jury.” Id. at 752, 90 S.Ct. 1463 (emphasis added). It is sometimes said that a plea is not “voluntary” if it is obtained through coercion that is “overbearing” to the defendant’s will. Brady, 397 U.S. at 750, 90 S.Ct. 1463. As noted in Miller v. Angliker, 848 F.2d 1312 (2d Cir.1988), a plea is valid “if it is not the product of actual or threatened physical harm, mental coercion overbearing the defendant’s will, or the defendant’s sheer inability to weigh his options rationally.” The words “overbearing” and “sheer inability” imply that some degree of “coercion” does not negate the validity of a guilty plea. Rather, forbidden is only “coercion” that is so overbearing to the defendant’s will that he or she cannot make a rational decision. Undue coercion may be “mental as well as physical” and may be the result of “subtle pressures” by the government on a defendant. Garrity v. New Jersey, 385 U.S. 493, 495, 87 S.Ct. 616, 17 L.Ed.2d 562 (1967). This flexible approach contradicts the absolute stand against coercion expressed by some courts. See United States v. Bell, 776 F.2d 965 (11th Cir.1985) (explaining the core requirements to a valid plea, including that the “plea must be free from coercion”); United States v. Dayton, 604 F.2d 931, 934 (5th Cir.1979) (stating that a guilty plea must be free from coercion); United States ex rel. Siebold v. Reincke, 362 F.2d 592, 593 (2d Cir.1966) (“A conviction will not be sustained if it rests upon a plea of guilty which is the result of coercion.”)- To be avoided, said one court, is “any semblance of coercion.” Euziere v. United States, 249 F.2d 293, 295 (10th Cir.1957) (“Fundamental standards of procedure in criminal cases require that a plea of guilty to the charge or charges contained in an indictment or information be entered freely, voluntarily, and without any semblance of coercion”). One court of appeals has realistically declared that a plea is involuntary only if it is obtained through “wrongful coercion.” United States v. Hernandez, 203 F.3d 614, 619 (9th Cir.2000). This implies that some coercion is implicit in the plea bargaining process and only “wrongful coercion”— however defined—invalidates a plea. That there are acceptable types and degrees of coercion is suggested by a holding that only some activities, such as judges’ participation in plea negotiations, are “inherently coercive.” See United States v. Barrett, 982 F.2d 193 (6th Cir.1992), citing United States ex rel. Elksnis v. Gilligan, 256 F.Supp. 244, 254 (S.D.N.Y.1966) (judges participation in plea negotiations is prohibited by Rule 11 of the Federal Rules of Criminal Procedure). An illegally coerced plea may be analogized to an illegally obtained confession. Some degree of coercion is present in many admissible confessions. One English court summed up the common law rule as follows: “A free and voluntary confession, is deserving of the highest credit, because it is presumed to flow from the strongest sense of guilt ... but a confession forced from the mind by the flattery of hope, or by the torture of fear, comes in so questionable a shape ... that no credit ought to be given to it; and therefore it is rejected.” King v. Warickshall, 1 Leach 262, 263-264, 168 Eng. Rep. 234, 235 (K.B.1783). The due process test of voluntariness in a confession requires the court to consider “the totality of all the surrounding circumstances— both the characteristics of the accused and the details of the interrogation ... [and] whether the defendant’s will was overborne.” Dickerson v. United States, 530 U.S. 428, 434, 120 S.Ct. 2326, 147 L.Ed.2d 405 (2000) (quoting Schneckloth v. Bustamonte, 412 U.S. 218, 226, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973)). Factors to be considered in determining whether the defendant rationally decided to confess include age, education, intelligence, length of detention and questioning, and physical and mental characteristics. See United States v. Hernandez, 893 F.Supp 952, 962 (D.Kan.1995), aff'd, United States v. Hernandez, 103 F.3d 145 (10th Cir.1996). Even wholly ratiocinatory economic creatures fully capable of adequately weighing the pros and cons of a prospective decision will take into account advantages and disadvantages offered by a prosecutor—in other words their coercive power. The carrot can be no less coercive than the stick—both impose pressure on decision. Cf. F.R.Cr.P. 11 (plea must be “voluntary and not the result of force or threats or of promises ”) (emphasis added). In addressing confessions as well as pleas, “voluntary” can not be defined by freedom from “coercion,” rather than by acknowledging appropriate levels of coercion. Determining whether a confession or plea is the product of illegitimate coercion requires consideration of the state of mind of the defendant as well as the techniques used for extracting agreement. Cf. Miller v. Fenton, 474 U.S. 104, 116, 106 S.Ct. 445, 88 L.Ed.2d 405 (1985) (“The admissibility of a confession turns as much on whether the techniques for extracting the statements, as applied to this suspect, are compatible with a system that presumes innocence and assures that a conviction will not be secured by inquisitorial means as on whether the defendant’s will was in fact overborne.”). More “coercion” may be allowable in the form of bargained for benefits in pleas than would be allowed in confessions because the plea is developed with counsel’s advice and with time to reflect and formalize the decision. A closer equivalence to “plea bargains” than confessions is suggested by contract law. Cf. Brady, supra, 397 U.S. at 752, 90 S.Ct. 1463 (plea bargains involve “mutuality of advantage”). The law recognizes that some degree of coercion exists in contract formation, but that the level of duress or threat may not be such that a party is “unfairly” (however defined) induced to enter into an agreement. Coercion is inherent in the bargaining process itself .... Every contract involves some kind of threat .... The problem is one of singling out those threats that amount to abuse of the bargaining process .... The rules on duress ... allow the injured party to undo the transaction by avoiding it. They seek to restore the parties to the positions in which they found themselves before they made the agreement. E. Allan Farnsworth, Contracts § 4.9 (Little, Brown & Co.1982). Yet there is no equivalence of a bargain between defendant and prosecutor and between entrepreneurs. For example, a party to a contract generally may not induce another party to agree by threat of criminal prosecution, whereas the nature of a plea negotiation is that not consenting necessarily results in further criminal prosecution. See Jamestown Farmers Elevator, Inc. v. General, 552 F.2d 1285, 1291 (8th Cir.1977) (“threats to institute criminal or regulatory proceedings ... made in order to secure another’s consent to an undeserved bargain for one’s own private benefit, may be sufficiently wrongful to constitute duress”); Farnsworth, supra, at § 4.20; United States v. Bethlehem Steel Corp., 315 U.S. 289, 300, 62 S.Ct. 581, 86 L.Ed. 855 (1942) (“The word duress implies feebleness on one side, overpowering strength on the other.”); see also N. Am. Rayon Corp. v. Comm’r of Internal Revenue, 12 F.3d 583, 589 (6th Cir.1993) (“Undue influence exists where a relationship of control exists between the contracting parties, and the stronger party influences the weaker party in a way that destroys the weaker party’s free will and substitutes for it the will of the stronger party.”); Commentary to § 175 of the Restatement (Second) of Contracts (a threat that would “arouse such fear as precludes a party from exercising free will and judgment or... [would] induce assent on the part of a brave man or a man of ordinary firmness” is voidable). In the context of plea negotiations, a defendant is likely to be in such a state of fear that he or she might in a civil action be considered “bereft of quality of mind essential for making a contract.” Rissman v. Rissman, 213 F.3d 381 (7th Cir.2000) (citation omitted). A prosecutor’s offer suggests “take this offer or I will use every resource at the government’s disposal to deprive you of your liberty for as long as possible.” This type of coercion is backed by the Federal Sentencing Guidelines. See Guidelines Manual § 3El.l(b)(2) (2001) (defendant’s sentence may be decreased by up to 3 levels by “timely notifying authorities of his intention to enter a plea of guilty, thereby permitting the government to avoid preparing for trial and permitting the court to allocate its resources efficiently”). “Coercion” and “voluntary” are, in short, vague terms of limited value in deciding whether to accept a plea. The critical question is whether the defendant was in a position to rationally weigh the advantages and disadvantages of the plea and whether a reasonable person in that position might make the same decision. This is a rather imprecise standard. In addition, certain tactics are considered incompatible with the etiquette of criminal justice, including overt threats and physical abuse. In accepting the plea the court must try to be as fully cognizant as practicable of the circumstances leading to the plea and of the nature and the background of the particular defendant, including age, education, social class, family pressures, and other relevant factors that might have affected the decision. The defense attorney’s position is critical for he or she provides defendant with the crutch of cool rationality. “Coercion” within limits is an available tool for prosecutors. It is appropriate so long as it does not shock the judicial conscience and does not depart substantially from commonly held beliefs of what is appropriate pressure for the government to apply to supposed miscreants. “Voluntary” under these circumstances means a capacity of the defendant and his counsel under the circumstances to rationally and fairly weigh the benefits of the plea against risks of not pleading. If “coerced” is too strong a word, one that leaves members of the justice system uncomfortable, many pleas are, to put the matter more politely, “induced” by strong promises of great value (e.g., a lesser term of incarceration, or life rather than death) — offers that, to paraphrase a famous movie line, most defendants cannot refuse. 2. Competency to Plead Voluntariness is comprised of at least two separate elements. First, the plea may not be “coerced” in the sense already described. “The agents of the State may not produce a [guilty] plea by actual or threatened physical harm or by mental coercion overbearing the will of the defendant.” Brady v. United States, 397 U.S. 742, 750, 90 S.Ct. 1463, 25 L.Ed.2d 747 (1970). Coercion is a fact-intensive inquiry which will depend on the particular case. Second, a defendant must be competent to plead. Saddler v. United States, 531 F.2d 83, 85 (2d Cir.1976). “If the judge, in response to his Rule 11 inquiries or because of information received from other sources, has reasonable grounds to doubt the defendant’s competence, he must refuse to accept the plea or defer acceptance pending a request for an examination of the defendant’s mental capacity.” Id. at 86. In making these two inquiries, courts are instructed to err on the side of caution and use “the 'utmost solicitude of which courts are capable in canvassing the matter with the accused to make sure [the defendant] has a full understanding of what the plea connotes and of its consequence.’ ” Id. at 85-86, citing Boykin v. Alabama, 395 U.S. 238, 243-44, 89 S.Ct. 1709, 23 L.Ed.2d 274 (1969). There is a lively debate among scholars over whether plea bargaining is systemically unreasonably coercive in nature. Compare, e.g., John H. Langbein, Torture and Plea Bargaining, 46 U. Chi. L.Rev. 3, 12-19 (1978) (plea bargaining is similar to medieval European torture), and Stephen J. Schulhofer, Plea Bargaining as Disaster, 101 Yale L.J.1979, 1980 (1992) (plea bargaining should be abolished), with Frank H. Easterbrook, Plea Bargaining as Compromise, 101 Yale L.J.1969, 1978 (1992) (plea bargaining is an efficient compromise and maintains defendant autonomy). Langbein’s view, while perhaps overly dramatic, is relevant to the present inquiry. He notes that: Like the Europeans of distant centuries who did employ [torture], we make it terribly costly for an accused to claim his right to the constitutional safeguard of trial. We threaten him with a materially increased sanction if he avails himself of his right and is thereafter convicted. This sentencing differential is what makes plea bargaining coercive. There is, of course, a difference between having your limbs crushed if you refuse to confess, or suffering some extra years of imprisonment if you refuse to confess, but the difference is of degree, not kind. Plea bargaining, like torture, is coercive. Langbein at 12-13 (emphasis added). Langbein further notes the irony in requiring that plea bargains be labeled “voluntary’ and “not coerced” before they can be approved by a court. “The plea agreement is the source of the coercion,” he notes, “and already embodies the involuntariness.” Id. at 14. His critique is rejected by the American system of criminal justice as it has evolved; it is now firmly bottomed on coercively induced pleas. 3. Changes in the Criminal Justice System When the Sentencing Guidelines were introduced, some judges expressed doubts about their constitutionality and fairness. See generally United States v. Yu, 1993 WL 497985 (S.D.N.Y. Nov.30, 1993) (Sweet, J.) (noting problems when guidelines were adopted, including the resignation of one district judge, findings of unconstitutionality by other judges, and a poll indicating that 92% of district judges opposed mandatory guidelines). One concern of some judges was that if it made no difference to the sentence under the guidelines whether a defendant was tried, or pleaded guilty, then the new process would lead to a substantial increase in the number of criminal trials. Defendants would, it was suggested, have no incentive to plead, since the sentence would be the same under the guidelines. In fact, just the opposite result occurred. The number of criminal jury trials has radically decreased since the guidelines were implemented. There are now relatively few criminal trials. In the meantime, the number of guilty pleas has increased dramatically. Cf. H.R.Rep. No. 93-1508 (1974), reprinted in 1974 USCCAN 7401, 7450 (hereinafter Speedy Trial Act History) (Letter from Attorney General opposing passage of the Speedy Trial Act, arguing that the legislation “would result in a decrease in the number of guilty pleas, since defendants would request jury trials with greater frequency”; in fact, as the statistics below show, the opposite occurred). One explanation for this largely unforeseen result is based on a series of related factors: The government often works out deals to avoid prosecution or drop more serious charges; 5K1 letters (which are issued, in accordance with U.S.S.G. § 5K.1 for defendants who cooperate with the government, and which grant a judge broad discretion to depart downward) are issued to many cooperating defendants; the government often does not contest findings of fact essential to lower sentences, such as “cooperation” or “minor role,” which can favor defendants; the government does not appeal many downward departures; and the government readily agrees to “safety valve” provisions (guidelines and statutes allowing for sentence reductions for certain defendants who cooperate with government, and for making inapplicable harsh statutory minimums). All of these benefits are available to the pleading defendant rather than to one taking the case to trial. Between safety valve and other reductions, defendants can often reduce their guidelines calculation by 6 or more levels, in effect cutting many sentences in half or more. See Guidelines Manual at 10 (“A change of six levels roughly doubles the sentence irrespective of the level at which one starts.”). The message to defendants is clear: “Don’t insist on your right to trial.” Since the sentencing judges’ ability to impose a sentence other than the one the government insists on under the guidelines is sharply reduced, the government’s power to induce pleas has been magnified. An officially published 1994 study by the Sentencing Commission, authored by a Sentencing Commissioner and a highly regarded law school professor, noted some of these dangers: Guideline factor bargaining remains common, especially with regard to the “role in the offense” guidelines provision. The common pattern is for the AUSA to agree to recommend (or not to oppose) a reduction for minor or minimal role when the facts do not support the recommendation.... The AUSA’s interpretation tends to be accepted by the court even if it is factually dubious .... An important vehicle for circumvention [of the guidelines] is the § 5K1.1 “substantial assistance” motion, which often is made on behalf of “sympathetic” defendants who have offered either minimal assistance, assistance that did not bear fruit in the way that § 5K requires, or no assistance at all. llene H. Nagel and Steven J Schulhofer, Plea Negotiations Under the Federal Sentencing Guidelines: An Empirical Examination of the Post-Mistretta Experience 22, 23 (1994) (hereinafter “Plea Negotiations Study”). The power of the prosecutor is particularly enhanced in connection with informers. If the government is satisfied that a putative defendant has made early admissions, particularly if accompanied by a promise to testify against other persons in the criminal enterprise, the government can provide a 5K.1 letter that permits probation for even the most heinous crimes accompanied by a waiver of any minimum sentence. See 18 U.S.C. § 3553(e) (court may impose sentence less than mandatory minimum in cases of substantial assistance); Guidelines Manual § 5K1.1, Application Note 1 (waiver of mandatory minimum sentence appropriate in some cases of assistance under § 5K1). The pressure on the informer to shade the truth to assist the government in obtaining convictions of others is enormous. Cf, e.g., Rothstein v. Corriere, 41 F.Supp.2d 381 (E.D.N.Y.1999) (case against false informant). While the government does all it can to ensure against false accusations, the use of informers does place greater pressure on informed against defendants to plead guilty. The court must guard particularly against pleas coerced in such situations. (In the instant case there was a strong hint of former employees of defendant informing on her.) Statistics available from the United States Sentencing Commission illustrate this trend to avoid the guidelines in order to induce pleas. The number of sentences within guideline range has dropped from 82% in 1989 to 64% in 2000. See Percent of Offenders Receiving Each Type of Departure 1996-2000, available online at http:// www.ussc.gov/ANNRPT/2000/fig-g.pdf, and Percent of Offenders Receiving Each Type of Departure 1989-1996, available online at http://www.ussc.gov/ annrpt/ 1996/fig-g.pdf. (The figures contained in these tables are included as Appendix 1 of this memorandum; they include a small number of upward departures which are not displayed on the graph because of space constraints). The majority of departures are granted for substantial assistance to authorities. Though the number of other departures has risen in recent years, it is still less than the number of departures for substantial assistance to authorities. Since substantial assistance departures require a letter from the government, the government has control over whether or not to grant a defendant the single most effective tool for receiving a sentence below the guideline range. No calculation is available of those “-within range” only because the charge has been reduced or the application has been manipulated or interpreted to avoid an apparent departure. See discussion below. There may, of course, be other explanations for the reduction in the relative number of trials besides utilization of the guidelines by the prosecutors to induce pleas — perhaps prosecutions are now those commenced only in the clearest cases of guilt, or perhaps now the mix of cases, emphasizing drug violations or illegal entry into the United States, has increased the ease of successful prosecution. At the same time that it essentially controls the sentence, the government can often block bail. In some cases this opposition is justified (such as for instances for many foreign nationals who import drugs and are likely to flee). Still, the combination of pretrial incarceration, plus higher periods of potential incarceration after conviction, when compared with a relatively easy plea and a known relatively short period of incarceration, create intense pressure on a defendant to plead. In most cases, this pressure does not result in a serious concern that the innocent are being found guilty. The 1994 Sentencing Commission study on plea negotiations conceded the existence of problems. Its study stated that “prosecutors exercise a considerable degree of sentencing discretion through charging and bargaining decisions. This discretion, if unchecked, has the potential to re-create the very disparities that the Sentencing Reform Act was intended to alleviate.” Plea Negotiations Study at 6. The report further noted that “circumvention” of the guidelines occurs, through bargaining by prosecutors in charging defendants. Id. at 7-8. It stated bluntly, “our research uncovered unequivocal evidence that bargaining and charging practices undercut the sentencing guidelines. There is simply no way to deny the existence of this problem in every jurisdiction we studied.” Id. at 17-18 (emphasis in original). Inappropriate use of minor role reductions and § 5K1 letters was found to lead to circumvention of the guidelines. Id. at 22-28. “The most important vehicle for guidelines evasion, post-Mistretta, is charge bargaining which leads to the dis.missal of readily provable counts,” id. at 23; prosecutors seeking cooperation from defendants sometimes use charge bargaining rather than § 5K1 reductions since § 5K1 “offers inadequate certainty [of reduced sentence].” Id. at 30. Statistics from the Administrative Office of the Federal Courts (some of which are available online through J-Net, at http:// jnet/hbrary/judfact/table3-5.htm) seem to indicate that this trend, towards guilty pleas and away from trials, was already underway before the implementation of the guidelines, but that the guidelines have accelerated it. They show that, between 1971 and 2001, the number of criminal defendants convicted in federal district courts rose from 78% of total defendants charged in 1971, and 75% of those charged in the next three years, to 90% of defendants charged in 2001. (Not all of the 10% of those charged but not convicted were acquitted. Dismissal may have taken place for a variety of reasons.) This ratio of convicted defendants to charged defendants has been driven up in part by large increases in the relative number of guilty pleas. The number of defendants convicted by guilty or nolo con-tendere plea has increased from 27,544 in 1971 to 64,402 in 2001. The total number of defendants convicted after jury trial dropped from 3,143 in 1971 (and an average of 3745 for the years 1971-75) to 2,294 in 2001. Significantly, the years immediately following the guidelines saw an increase in the number of defendants convicted after jury trials, from 4,191 in 1988, to 4,640 in 1989, to 4,851 in 1990, to over 5000 each in 1991 and 1992. Since 1992, the number of defendants convicted after jury trials has decreased in every year except one, and the ten years of decreases have halved the number of defendants convicted after jury trials. Overall, the percent of convictions resulting from guilty plea has increased from 86% in 1971 (27,544 pleas in a total of 32,103 convictions) to 95% in 2001 (64,402 pleas in a total of 67,731 convictions). The statistics for the district court for the Eastern District of New York are similar. In 1971, the Eastern District had a total of 147 (15.5% of total) defendants convicted after jury trials, and 28(3%) convicted after bench trials, with 773 (81.5% of total) convicted by plea. By 2001, there were only 48 defendants convicted by jury trial and a single defendant convicted by bench trial (3.5% of total) versus 1336 (96.5% of total) convicted by plea. These numbers should not surprise anyone familiar with the current criminal law system. They are available in part in the Sentencing Commission’s Sourcebook. See 2000 Sourcebook of Federal Sentencing Statistics, Figure C (“Guilty Pleas and Trial Rates”). (The Sourcebook’s numbers for percent of pleas differ slightly from those provided by the Accounting office, varying as much 0.5%; the reason for this slight variation is unknown to the court.) Numbers of pleas and trials for state courts, while not readily available, are likely to be similar to those for Federal courts. Cf. Plea Negotiations Study at 11 (in Minnesota, “the power of prosecutors unquestionably increased” after state sentencing guidelines were adopted; much of the shift is due to increased “charge bargaining” over the nature of charges); id. at 12-13 (adoption of Pennsylvania guidelines increased conviction rate; “plea practices in urban areas were altered”). The Advisory Committee Note to Rule 11 of the Federal Rules of Criminal Procedure discusses the issue in passing: Although reliable statistical information is limited, one recent estimate indicated that guilty pleas account for the disposition of as many as 95% of all criminal cases. ABA Standards Relating to Pleas of Guilty, pp. 1-2 (Approved Draft, 1968). A substantial number of these are the result of plea discussions. The President’s Commission on Law Enforcement and Administration of Justice, Task Force Report: The Courts 9 (1967); D. Newman, Conviction: The Determination of Guilt or Innocence Without Trial 3 (1966); L. Weinreb, Criminal Process 437 (1969); Note, Guilty Plea Bargaining: Compromises By Prosecutors To Secure Guilty Pleas, 112 U.Pa.L.Rev. 865 (1964).... Administratively, the criminal justice system has come to depend upon pleas of guilty and, hence, upon plea discussions. See, e.g., President’s Commission on Law Enforcement and Administration of Justice, Task Force Report. The Courts 9 (1967); Note, Guilty Plea Bargaining: Compromises By Prosecutors To Secure Guilty Pleas, 112 U.Pa.L.Rev. 865 (1964). F.R.Cr.P. 11, Advisory Committee Notes to 1974 Amendments. Following this trend, the number of trials (including criminal jury trials) handled by each Federal judge annually has dropped steadily. In 1992, the average judge handled 32 trials (bench and jury, civil plus criminal) in a year; there were an average of 9 criminal jury trials per year. By 2001, the average judge handled 20 trials in a year; there were an average of 5 criminal jury trials per year. During the same time, the annual number of criminal cases filed per judge rose, from 54 in 1992 to 77 in 2001. The statistics are directly available online (except for the number of criminal jury trials per year); see Judicial Caseload Profile, available online at www.uscourts.gov/cgi-bin/ cmsd2001.pl, and Judicial Caseload Profile, available online at www.useourts.gov/cgi-bin/cms.pl. The number of criminal jury trials is derived from the number of judges per year, see citations supra, and the number of criminal jury trials which is available at http://jneVlibrary/judfact/table4-3.htm. (These figures appear compatible; for example, the numbers of total trials available at the J-Net site corresponds exactly with the numbers of trials per judge available in the CMS statistics). Over the past 30 years, the number and percentage of defendants convicted and sentenced in Federal courts has increased dramatically, both nationally and within this district. The following charts illustrate these patterns. Corresponding changes to the number of defendants acquitted further illustrate this trend. These numbers are related to the increase in guilty pleas taken, both in total number and as a percent of convictions. As both the total and relative numbers of guilty pleas have increased, the number of trials has decreased, both in total and relative terms, while the number of criminal cases per judge has increased. As the above chart demonstrates, while Federal judges have seen an almost 50% rise in criminal cases per judge, from 54 to 77 per year in the last 10 years, they have also observed the total number of trials (bench and jury, civil and criminal) drop by more than a third, from 32 to 20; the number of criminal jury trials has dropped by nearly half. (For display purposes, this chart is rounded to the nearest whole numbers (9 trials per year for 1992 and 5 for 2001); calculations of decrease are not helpful with these rough numbers. An exact calculation can be made using the more precise numbers of 8.84 for 1992 and 5.05 for 2001, accurately showing a 43% decrease in the number of criminal jury trials per judge since 1992.) When examining the criminal justice system in its report on the Speedy Trial Act the House Judiciary Committee noted the danger signal flashed by our increased reliance on pleas. It declared: Whether the negotiated plea is a desirable element within the system or one of the basic causes of delay and court-clogging is another question. The National Advisory Commission in its Courts report found that plea bargaining constitutes a triple danger to the system: (1) Danger to the Defendant’s Rights — A survey of more that 3,400 criminal justice practitioners in four states — California, Michigan, New Jersey, and Texas — -revealed that 61 percent of those polled agreed that it was probable or somewhat probable that most defense attorneys engage in plea bargaining primarily to expedite the movement of cases. Furthermore, 8 percent agreed that it was probable or somewhat probably that most defense attorneys in plea bargaining negotiations pressure clients into entering a plea that the client feels is unsatisfactory. (2) Danger to Court Administration — Very simply, the Commission found that plea bargaining resulted in the need to pull cases out of the process — sometimes on the morning of trial — making efficient scheduling of cases difficult or impossible. Thus, plea bargaining makes it difficult to use judicial and prosecutorial time effectively. (3)Danger to Society’s Need for Protection — -The conclusion of the commission in this regard is that, because defendants are often dealt with less severely than might normally be the case, plea bargaining results in leniency that reduces the deterrent effect of the law and may have a less direct effect on corrections programs. Speedy Trial Act History at 7412 (emphasis supplied). Because of these factors, the Commission recommended the prohibition of plea bargaining. Id. Rejecting this radical recommendation, the House Committee suggested only that the report deserved “weight.” The Sentencing Commission found that the result of “charge bargaining” by prosecutors was that “the amount of [sentence reduction] is taken out of the hands of the judge, where the Guidelines intended to place it, and left more under control of the prosecutor, where neither the Congress nor the Commission intended the discretion to be.” Plea Negotiations Study at 30-31. Most pleas of guilty are fully justified. There is often clear evidence of guilt. Often the defendant agrees on his or her guilt at the time of arrest. The catharsis of prompt acknowledgment of wrongdoing and its potential first step on the road to rehabilitation should not be underestimated as a value of the truly “voluntary” plea. Occasionally, as in the case at hand, the coercive nature of the system is such as to give serious cause for concern that we are utilizing excessive coercion to avoid trials. Yet, there is no injustice on any appreciable scale through conviction of the innocent that can be demonstrated. Neither prosecution nor defense, nor the trial nor appellate courts, seem blameworthy. Nevertheless, the end result of the “new criminal justice system” in almost eliminating trials is disquieting. Such massive shifts may in practice require development of new judicial techniques, ethical rules or attitudes. Cf, e.g. Mae C. Quinn, Whose Team Am I on Anyway?, 26 Rev. L. Soc. Ch. 37, 38 (2001) (noting new model of “teamwork” between judge, prosecutor, and defense in drug cases, requiring change in attitude of defense counsel towards client and court). C. Bail 1. Law Governing Bail Bail in Federal courts is governed by section 3142 of Title 18 of the United States Code. Section 3142 establishes a general policy of release of prisoners pending trial. Prisoners may be released on their own recognizance (subsection (b)) or on conditions (subsection (c)). Only if, after conducting hearings, the judge decides that release presents a danger to the security of the public or that “no condition or combination of conditions will reasonably assure that appearance of the person as required,” may the defendant be detained without bail. 18 U.S.C. § 3142(e). This provision codifies the longstanding norm in our justice system that bad is the rule, and that few cases are to be exceptions. The Constitution itself prohibits the use of excessive bail. U.S. Const. Am. VIII. “At common law in England pretrial release on bail was a matter within the discretion of judges for all defendants.” United States v. Melendez-Carrion, 790 F.2d 984, 997 (giving history of bail under common law, citing authorities including Blackstone). This common law tradition was continued by the Judiciary Act of 1789, creating a right to bail in all but capital cases, where the court had the option of non-release. Jud. Act of 1789, Ch. 20, 1 Stat. 91 (1789), cited in Melendez-Carrion, supra. Later changes in the law made all prisoners subject to bail requirements. The basic theory of freedom before conviction, however, never changed. As the Supreme Court reiterated in Stack v. Boyle, 342 U.S. 1, 72 S.Ct. 1, 96 L.Ed. 3 (1951): From the passage of the Judiciary Act of 1789, 1 Stat. 73, 91, to the present Federal Rules of Criminal Procedure, Rule 46(a)(1), 18 U.S.C.A., federal law has unequivocally provided that a person arrested for a non-capital offense shall be admitted to bail. This traditional right to freedom before conviction permits the unhampered preparation of a defense, and serves to prevent the infliction of punishment prior to conviction. Unless this right to bail before trial is preserved, the presumption of innocence, secured only after centuries of struggle, would lose its meaning. Id. at 4, 72 S.Ct. 1 (Emphasis added). 2. History of Bail Reform Despite these protections, many poor defendants were often unable to make bail. A disparity based upon social status led to widespread dissatisfaction. Public interest organizations urged increased release before trial. There were several reasons why bail reform was urged. A primary concern was that expressed by the Supreme Court earlier in Stack, that defendants who are held in pretrial custody are less able to defend themselves. Stack, 342 U.S. at 4, 72 S.Ct. 1. This correlation was shown in several studies prior to passage of the Bail Reform Act in 1964. See, e.g., Anne Rankin, The Effect of Pretrial Detention, 39 N.Y.U. L.Rev. 641 (1964). Ms. Rankin, a Research Associate from the Vera Institute, relied upon data from New York. She pointed out that “previous studies of bail have indicated that an accused who has been detained in jail between his arraignment and the final adjudication of his case is more likely to receive a conviction or a jail sentence than an accused who has been free on bail.” Id. at 641. She used standard statistical techniques to attempt to neutralize the effect of other variables. An initial reading of the statistics showed that for the period in question, sixty-four percent of non-bailed defendants were sentenced to prison time, while only seventeen percent of bailed defendants were sentenced to prison. Id. at 643. This difference of forty-seven percent was largely attributable, the study concluded, to the different decisions made on bail. Factors one might think would cause a difference, such as the race of the defendant and the type of offense charged, were not statistically significant. Id. at 644. Rankin identified five factors which did have a statistical correlation with disposition: defendant’s previous