By way of background, most criminal tax cases, as most criminal cases generally, are resolved by plea agreement. (See the addendum below on a recent article on statistics dealing with the role of pleas in the federal criminal system.) For example, assume that potential client A comes into your office the day after he was indicted for several tax crimes. A announces that he came to you because you had the reputation of being the best criminal tax lawyer in the universe (you modestly but not totally candidly disavow that reputation). A then outlines his cryptic view of why he is innocent. He then asks what are the chances of you obtaining an acquittal for him. All you know is his cryptic account which or may not be a fair representation or summary of the facts, but that cryptic account proclaims his complete innocence, at least on the willfulness element of the tax crimes charged because he says he is innocent. So, you remind him that he is asking you to state conclusions based on cryptic facts which may or may not be true and which you have not investigated. On that basis, you advise first that, if the cryptic statement is a fair representation of the case that will be presented at trial, then he has a very good chance of being acquitted. You then state that your experience is that such cryptic initial accounts generally are too cryptic for anyone to feel comfortable that that is the way it will play out at trial. You then state that, given the highly selective systemic selection of criminal tax cases, culminating in DOJ Tax and AUSA review before indictment, the facts may well not play out that nicely at trial. You then tell him, that given that selectivity, the posted rates of conviction in tax cases are very high -- exceeding 90% (maybe, see my several other postings on the conviction rates in tax cases). You then tell him, based on that statistic alone, and discounting his cryptic proclamation of innocence, there is a 90+% chance he will be convicted. You finally tell him there is a systemic preference in the federal criminal system generally and in the tax crimes subset of that system to resolve cases by plea -- indeed a defendant is given a substantial benefit in the Sentencing Guidelines by resolving the case by plea. The combination of likely conviction and the benefit of pleading for a reduced sentence is a powerful incentive to plead. (Indeed as others have noted, it may be so powerful in some cases that the innocent plead, provided that they can clear the hurdle of allocution of guilt.)
Now, with that background, the defendant in Davila was unhappy with his attorney who, apparently with more facts in hand as to how the trial would play out, advised the Davila to accept the plea proffered by the prosecutor. Davila interpreted that recommendation as being a reflection of the fact that the attorney had no defensive strategy. That is probably a fair lay interpretation, but criminal defense lawyer would characterize the recommendation as a conclusion that the defendant almost certainly would be convicted. Davila's strategy then was to request new counsel. An ex parte hearing or meeting on that request was held with the U.S. Magistrate Judge. The prosecutor was not present (not clear why, but that was a no-no.). During the course of the hearing or meeting, the Magistrate Judge advised that defendant that he would not get another court-appointed attorney and that his best course, given the strength of the Government's case, was to accept the plea. Davila was not convinced. Time passed. Finally, he became convinced and pled with a full allocution saying that he was guilty of the crime to which he pled and that nothing had been promised him in return, etc., etc.
The issue was whether the Magistrate Judge inappropriate advised or suggested that the defendant plead, despite a prohibition in the Federal Rules of Criminal Procedure that prevents judges from participation in such plea discussion. I think that the key point for readers is discernible from the "Syllabus" which I quote in its entirety:
Federal Rule of Criminal Procedure 11 governs guilty pleas. Rule 11(c)(1) instructs that “[t]he court must not participate in [plea] discussions,” and Rule 11(h) states that a “variance from the requirements of th[e] rule is harmless error if it does not affect substantial rights.” Rule 52(a), which covers trial court errors generally, similarly prescribes: “Any error . . . that does not affect substantial rights must be disregarded.”
Respondent Davila, while under indictment on multiple tax fraud charges, wrote to the District Court, expressing dissatisfaction with his court-appointed attorney. Complaining that his attorney offered no defensive strategy, but simply advised him to plead guilty, Davila requested new counsel. A Magistrate Judge held an in camera hearing at which Davila and his attorney, but no representative of the United States, appeared. At the hearing, the Magistrate Judge told Davila that he would not get another court-appointed attorney and that his best course, given the strength of the Government’s case, was to plead guilty. More than three months later, Davila pleaded guilty to a conspiracy charge in exchange for dismissal of 33 other charges. He stated under oath before a U. S. District Judge that he had not been forced or pressured to enter the plea, and he did not mention the in camera hearing before the Magistrate Judge. Prior to sentencing, however, Davila moved to vacate his plea and dismiss the indictment, asserting that he had entered the plea for a “strategic” reason, i.e., to force the Government to acknowledge errors in the indictment. Finding that Davila’s plea had been knowing and voluntary, the District Judge denied the motion. Again, Davila said nothing of the in camera hearing conducted by the Magistrate Judge. On appeal, the Eleventh Circuit, following Circuit precedent, held that the Magistrate Judge’s violation of Rule 11(c)(1) required automatic vacatur of Davila’s guilty plea, obviating any need to inquire whether the error was prejudicial.
Held: Under Rule 11(h), vacatur of the plea is not in order if the record shows no prejudice to Davila’s decision to plead guilty. Pp. 7-14.
(a) Rule 11(c)(1)’s prohibition of judicial involvement in plea discussions was included in the 1974 Amendment to the Rule out of concern that a defendant might be induced to plead guilty rather than risk antagonizing the judge who would preside at trial. Rule 11(h) was added in the 1983 Amendment to make clear that Rule 11 errors are not excepted from Rule 52(a)’s harmless-error inquiry. Rule 52 also states, in subsection (b), that a “plain error that affects substantial rights may be considered even though it was not brought to the [trial] court’s attention.” When Rule 52(a) governs, the prosecution has the burden of showing harmlessness, but when Rule 52(b) controls, the defendant must show that the error affects substantial rights. See United States v. Vonn, 535 U. S. 55, 62, 122 S. Ct. 1043, 152 L. Ed. 2d 90.
As clarified in Vonn and United States v. Dominguez Benitez, 542 U. S. 74, 124 S. Ct. 2333, 159 L. Ed. 2d 157, Rule 11 error may be of the Rule 52(a) type or the Rule 52(b) kind, depending on when the error was raised. In Vonn, the judge who conducted the plea hearing failed to inform the defendant, as required by Rule 11(c)(3), that he would have “the right to the assistance of counsel” if he proceeded to trial. The defendant first objected to the omission on appeal. This Court held that “a silent defendant has the burden to satisfy [Rule 52(b)’s] plain-error rule.” 535 U. S., at 59, 122 S. Ct. 1043, 152 L. Ed. 2d 90. In Dominguez Benitez, the error first raised on appeal was failure to warn the defendant, as Rule 11(c)(3)(B) instructs, that a plea could not be withdrawn even if the sentence imposed was higher than the plea-bargained sentence recommendation. The Court again held that Rule 52(b) controlled, and prescribed the standard a defendant silent until appeal must meet to show “plain error,” namely, “a reasonable probability that, but for the [Rule 11] error, he would not have entered the plea.” 542 U. S., at 83, 124 S. Ct. 2333, 159 L. Ed. 2d 157. Pp. 7-9.
(b) Here, the Magistrate Judge plainly violated Rule 11(c)(1) by exhorting Davila to plead guilty. Davila contends that automatic vacatur, while inappropriate for most Rule 11 violations, should attend conduct banned by Rule 11(c)(1). He distinguishes plea-colloquy omissions, i.e., errors of the kind involved in Vonn and Dominguez Benitez, from pre-plea exhortations to admit guilt. The former come into play after a defendant has decided to plead guilty, the latter, before a defendant has decided to plead guilty or to stand trial. Nothing in Rule 11’s text, however, indicates that the ban on judicial involvement in plea discussions, if dishonored, demands automatic vacatur without regard to case-specific circumstances. Nor does the Advisory Committee commentary single out any Rule 11 instruction as more basic than others. And Rule 11(h), specifically designed to stop automatic vacaturs, calls for across-the-board application of the harmless-error prescription (or, absent prompt objection, the plain-error rule).
Rule 11(c)(1) was adopted as a prophylactic measure, not one impelled by the Due Process Clause or any other constitutional requirement. Thus, violation of the Rule does not belong in the highly exceptional category of structural errors—e.g., denial of counsel of choice or denial of a public trial—that trigger automatic reversal because they undermine the fairness of the entire criminal proceeding. United States v. Marcus, 560 U. S. 258, ___, 130 S. Ct. 2159, 176 L. Ed. 2d 1012. Instead, in assessing Rule 11 errors, a reviewing court must take account of all that transpired in the trial court. Had Davila’s guilty plea followed soon after the Magistrate Judge’s comments, the automatic-vacatur rule would have remained erroneous. The Court of Appeals’ mistake in that regard, however, might have been inconsequential, for the Magistrate Judge’s exhortations, if they immediately elicited a plea, would likely have qualified as prejudicial. Here, however, three months distanced the in camera meeting conducted by the Magistrate Judge from Davila’s appearance before the District Judge who examined and accepted his guilty plea after an exemplary Rule 11 colloquy, at which Davila had the opportunity to raise any questions he might have about matters relating to his plea. The Court of Appeals, therefore, should not have assessed the Magistrate Judge’s comments in isolation. Instead, it should have considered, in light of the full record, whether it was reasonably probable that, but for the Magistrate Judge’s comments, Davila would have exercised his right to go to trial. Pp. 10-14.
(c) The Court of Appeals, having concluded that the Magistrate Judge’s comments violated Rule 11(c)(1), cut off further consideration. It did not engage in a full-record assessment of the particular facts of Davila’s case or the case-specific arguments raised by the parties, including the Government’s assertion that Davila was not prejudiced by the Magistrate Judge’s comments, and Davila’s contention that the extraordinary circumstances his case presents should allow his claim to be judged under Rule 52(a)’s harmless-error standard rather than Rule 52(b)’s plain-error standard. The Court decides only that the automatic-vacatur rule is incompatible with Rule 11(h) and leaves all remaining issues to be addressed on remand. P. 14.
664 F. 3d 1355, vacated and remanded.
The Wall Street Journal and the Washington Post recently presented some sobering statistics on the high rate of uncontested federal convictions. The editorial in the Post (Oct. 4, 2012) both copied and adopted the Journal's figures: "The Wall Street Journal reported recently that 97 percent of cases the Justice Department prosecuted last year ended with guilty pleas."
Both papers were misreading Justice Kennedy's opinion for the Court in Missouri v. Frye, 132 S. Ct. 1399 (2012). The opinion observed that guilty pleas account for "97 percent of federal convictions," not "97 percent of cases the Justice Department prosecuted" as the Post and Journal said.
The Journal's statistic overlooks the substantial number of cases that are not prosecuted to conviction at all. The Bureau of Justice Statistics Sourcebook for 2009, to which Justice Kennedy's opinion cites for the guilty plea figures, shows that guilty pleas were secured in only about 88 percent of the cases disposed of in 2009 (83,707 out of 95,206), not 97 percent. In the remaining 12 percent of the federal prosecutions, the Sourcebook's Table shows about 3 percent of the total number of defendants were found guilty at a trial, and about one-half of 1 percent won acquittals at trial from the jury or the court. That leaves about 8V2 [sic - apparently 8 1/2] percent of the cases that are listed in the Statistical Table as Dismissed.
But the Department of Justice's published statistics give no breakdown of the grounds on which these numerous defendants (about 1 in every 12 charged, or 8,408 of the 95,206 dispositions in 2009) secured dismissal of their cases.
It would be of great interest, especially to the defense bar, to have a breakdown of the large "case dismissed" category. For example, among the 8,408 dismissals in 2009, what number or percentage of the dismissals were secured through a successful defense motion to dismiss the indictment for (1) failure to charge an offense against the United States, (2) discriminatory or otherwise improper grand jury selection practices, (3) violation of grand jury procedural rules, (4) prosecutorial misconduct, or (5) other defects in the indictment or prosecution? And how many of the cases were dismissed on motion of the government under Fed. Rule Crim. P. 48 because (1) the prosecution believed it lacked sufficient credible evidence to win a conviction at trial, (2) defense counsel presented persuasive argument, or new exculpatory or favorable evidence, to the counsel for the government, (3) the government moved to dismiss the indictment pursuant to a plea bargain, or in return for supplying testimony against others, or (4) the defendant died or had otherwise become unavailable?I probably will return to this statistics subject again. I have emailed a request to Professor Weinberg to discuss the issue and have not yet heard back. I would definitely like his feedback on my concern about the statistics in the criminal tax prosecution arena.
Keep in mind also that Professor Weinberg's comments relate to federal criminal cases generally and not just the tax crimes subset. The immediate question his comments raise is whether the tax crimes case selection process is that much more rigorous than the overall federal criminal case selection process?