192 F.3d 179 (1st Cir. 1999)
UNITED STATES OF AMERICA, Appellee,
JOSE E. ALEGRIA, Defendant, Appellant.
United States Court of Appeals for the First Circuit.
Heard Aug. 6, 1999.
Decided Sept. 30, 1999.
APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF PUERTO RICO
Alan M. Dershowitz, with whom Nathan Z. Dershowitz, Amy Adelson, and Dershowitz & Eiger, P.C. were on brief, for appellant.
Jorge E. Vega-Pacheco, Assistant United States Attorney, with whom Guillermo Gil, United States Attorney, and Nelson Perez-Sosa, Assistant United States Attorney, were on brief, for appellee.
Before Selya, Boudin and Lipez, Circuit Judges.
SELYA, Circuit Judge.
This appeal requires us, inter alia, to explore the circumstances in which the government may be compelled to move for a downward departure under USSG §5K1.1. We conclude that the district court did not err either in refusing to force the government to take such action or in any other material respect. Consequently, we affirm.
* A federal grand jury indicted defendant-appellant Jose E. Alegra on sixteen counts of filing false statements with financial institutions, 18 U.S.C. § 1014, and bank fraud, 18 U.S.C. § 1344. He entered into a plea agreement with the government (the Agreement), pled guilty to all charges, and met twice with government agents pursuant to a promise to cooperate. We have no detailed account of these debriefing sessions, but the appellant states in a declaration (filed below in connection with his motion for an evidentiary hearing) that he furnished the government with whatever information he possessed concerning wrongdoing at the financial institutions with which he was associated.
Despite the appellant's cooperation, the prosecutor elected not to file a downward departure motion. The appellant asserted that the prosecutor's decision contravened the Agreement and, in the bargain, violated due process. The sentencing court rejected these animadversions, see United States v. Alegra, 3 F. Supp. 2d 151 (D.P.R. 1998), denied the appellant's motion for an evidentiary hearing, and proceeded to impose a 30-month incarcerative sentence. In this forum, the appellant continues to press his claim that the government wrongly refused to file a downward departure motion and embellishes it with a challenge to the lower court's calculation of his guideline sentencing range.
We start with the appellant's major premise: that the government obligated itself to file a downward departure motion by virtue of promises it made during the negotiations that led up to the execution of the Agreement and in the Agreement itself. For argument's sake, we take the facts from the appellant's declaration.1
After the indictment was returned and the appellant entered a "not guilty" plea, the parties began discussing the possibility of a plea bargain. In his declaration, the appellant states that he had misgivings about whether the United States Attorney's office would reward cooperation with a favorable sentencing recommendation (he traces these misgivings to a previous case in which the United States Attorney allegedly made similar overtures to another bank executive, but subsequently reneged), and therefore arranged to meet personally with Guillermo Gil, the United States Attorney for the District of Puerto Rico, prior to settling upon a course of action. According to the appellant, Gil assured him (in the presence of his then-counsel) that if he would "tell the truth, be available, and cooperate," the government would move for a departure under USSG §5K1.1 (permitting a sentencing court to depart downward on the prosecution's motion, based on a defendant's "substantial assistance"). The appellant asserts that this specific representation persuaded him to sign the Agreement and change his plea. Hence, he asks that we hold the government to Gil's word.
As a general rule, nothing precludes a prosecutor from bargaining away something over which he has discretion in return for promises extracted from a criminal defendant. See United States v. Doe, 170 F.3d 223, 226 (1st Cir. 1999); United States v. Hernandez, 17 F.3d 78, 82 (5th Cir. 1994). Relatedly, a binding prosecutorial representation that is accepted by a defendant and becomes the basis for a change of plea must be performed. See Santobello v. New York, 404 U.S. 257, 262 (1971) (holding that "when a plea rests in any significant degree on a promise or agreement of the prosecutor, so that it can be said to be part of the inducement or consideration, [the] promise must be fulfilled"). In the appellant's view, these uncontroversial axioms carry the day.
But this conclusion depends entirely on the assumption that what Gil allegedly said has legal force -- and that assumption stands on shaky ground because the Agreement, which purports to encompass the sum and substance of the arrangement between the parties, was signed after Gil allegedly made the crucial representation and contains no reference to it. The essential and logically prior question, then, is whether the representation, even if made, survives execution of the Agreement.
* Courts customarily treat plea agreements, for purposes of construction, more or less in the same manner as they do contracts. See United States v. Atwood, 963 F.2d 476, 479 (1st Cir. 1992); United States v. Anderson, 921 F.2d 335, 337-38 (1st Cir. 1990). We say "more or less" because this analogy has its limitations. See United States v. Hogan, 862 F.2d 386, 388 (1st Cir. 1988) (observing that plea agreements are similar to commercial contracts, but only "in certain respects"). Thus, although contract law supplies a useful reference point for construing plea agreements in federal criminal cases, such agreements are not governed by the law of contracts. See United States v. Kelly, 18 F.3d 612, 616 (8th Cir. 1994).
If a plea agreement unambiguously resolves an issue, that usually ends the judicial inquiry. See id.; Anderson, 921 F.2d at 338. If, however, a plea agreement lacks clarity or is manifestly incomplete, the need to disambiguate may justify resort to supplementary evidence or other interpretive aids.2 See Anderson, 921 F.2d at 338. We examine the text of the Agreement in light of this dichotomy.
The appellant calls our attention to paragraph 8 of the Agreement, which embodies his pledge to "cooperate fully and truthfully with the United States." After spelling out the elements of this cooperation -- which include the typical assurances that the appellant will remain available for debriefing, appear as a witness, speak truthfully, provide documents, and so forth -- the paragraph explains that he is not expected to "make a case" against anyone. This is a shorthand way of saying that the government's obligations under the Agreement are not conditioned upon the achievement of any particular objective (e.g., the conviction of some other person), but, in the language of the Agreement, "only upon [Alegra] providing full, complete and truthful cooperation." The appellant maintains that this phraseology imports the United States Attorney's oral representation into the Agreement. We think that this is a more ambitious reading of the passage than either the text or the surrounding circumstances allow.
USSG §5K1.1 provides the background understanding against which the parties signed the Agreement and through which their arguments must be filtered. See United States v. Huang, 178 F.3d 184, 187-89 (3d Cir. 1999). The appellant's construction contemplates an equivalency between "full, complete and truthful cooperation," on the one hand, and "substantial assistance," on the other. But the language and structure of section 5K1.1 belie the idea that "full, complete and truthful cooperation" necessarily constitutes "substantial assistance." The guideline suggests five non-exclusive factors that a court should consider when deciding whether it will grant a prosecutor's motion for a downward departure predicated on a defendant's substantial assistance. See USSG §5K1.1(a)(1)-(5). Full, complete and truthful cooperation corresponds to only one of these five factors. See id. §5K1.1(a)(2). The others include things well beyond the purview of cooperation per se, such as the significance and utility of the information provided, id. §5K1.1(a)(1), the nature and extent of the defendant's assistance, id. §5K1.1(a)(3), and the timeliness of the proffer, id. §5K1.1(a)(5). In short, full, complete and truthful cooperation, in and of itself, is not coextensive with the substantial assistance of which the sentencing guidelines speak.
In the case at bar, the Agreement, read as a whole, plainly was meant to be understood in terms of the general approach limned in section 5K1.1. Although conditioned on the appellant's conformance with paragraph 8, nothing in the text of the Agreement suggests that the parties agreed either to collapsing the substantial assistance determination into the relatively narrow confines of paragraph 8 or to some other special definition of substantial assistance. This point is made pellucid by paragraph 11, which memorializes the appellant's express agreement that "the United States' decision whether to file a motion based on 'substantial assistance' as that phrase is used in Rule 35(b) of the Federal Rules of Criminal Procedure and Section 5K1.1 of the Sentencing Guidelines and Policy Statements . . . rests in the sole discretion of the United States," and further provides that disputes about that decision will not be referred to the district court.
The obvious implication of this explicit reference to section 5K1.1 is that the prosecutor will take into account all the factors delineated in that guideline when determining whether to move for a downward departure -- and those factors, as we have noted, go well beyond full, complete and truthful cooperation. In this way, the Agreement makes it quite clear that compliance with the covenants contained in paragraph 8 constitutes a necessary, but not an independently sufficient, precondition to the filing of a section 5K1.1 motion.3 Moreover, the reference in paragraph 11 to Fed. R. Crim. P. 35(b) bolsters, rather than weakens, this conclusion: with regard to the meaning of "substantial assistance," Rule 35(b) and USSG §5K1.1 are birds of a feather. See United States v. Gangi, 45 F.3d 28, 30 (2d Cir. 1995).
The appellant's exhortation that our decision in Doe, 170 F.3d 223, points in a different direction is easily dispatched. Although the Doe court spoke of internal "tension" within the contours of a plea agreement, that tension related to a completely different problem: the plea agreement purported to preserve the government's absolute discretion in regard to section 5K1.1 motions, but at the same time stated that "the defendant's failure to 'make a case' shall not relieve the government of exercising its discretion" under section 5K1.1. Id. at 226. The government decided not to file a downward departure motion because Doe's assistance "came too late." Id. In that situation, we suggested that the government's performance arguably conflicted with the assurance contained in the plea agreement. Concerned that the government may have declined to file a section 5K1.1 motion because Doe's help had come "too late" to permit it to "make a case" against a third party, we found some tension between what the government said it would do (i.e., not peg the substantial assistance determination on whether Doe had made a case against someone) and what it actually did. See id.
There is a critical difference between this case and Doe. The Agreement sub judice does not tie the government's exercise of its section 5K1.1 discretion to whether the defendant has (or has not) made a case against a third party, and thus does not contain the language that caused the contretemps in Doe. What is more, Doe does not in any way intimate that the mere inclusion of a cooperation clause in a plea agreement somehow circumscribes ex proprio vigore the government's discretion anent the filing of such motions. Indeed, Doe never argued (as Alegra does) that cooperation, if rendered, mandates the filing of a section 5K1.1 motion.
We have said enough on this score. The short of it is that the concepts of "full, complete and truthful cooperation" and "substantial assistance" are neither congruent nor interchangeable, and the plain text of paragraph 11 refutes the appellant's contention that the parties expressly modified this basic understanding. Consequently, the government's election not to file a section 5K1.1 motion did not violate the stated terms of the Agreement.
The appellant has a fallback position. He invites us to supplement the Agreement by engrafting onto it the oral representation allegedly made by the United States Attorney during the pre-plea negotiations. We decline the invitation.
The appellant's position flies in the teeth of paragraph 22 of the Agreement, which states flatly that the written document constitutes the complete agreement between the parties and that the "United States has made no promises or representations except as set forth in writing in this plea agreement and deny [sic] the existence of any other terms and conditions not stated herein." Where, as here, an unambiguous plea agreement contains an unqualified integration clause, it normally should be enforced according to its tenor. That means, of course, that an inquiring court should construe the written document within its four corners, "unfestooned with covenants the parties did not see fit to mention." Anderson, 921 F.2d at 338.
In United States v. Burns, 160 F.3d 82, 83 (1st Cir. 1998), we decisively rejected a similar attempt by a defendant to read into a written plea agreement an implied constraint on the government. Burns, in the course of appealing the district court's enhancement of his sentence under a guideline provision, argued that a clause in the plea agreement which restrained the government from recommending such an increase "at sentencing" implied a duty not to oppose Burns's effort to set aside the increase on appeal. See id. at 83. We rejected this argument, explaining that the government's promise simply did not go so far. See id. Moreover, we admonished that "significant plea-agreement terms should be stated explicitly and unambiguously." Id. Alegra, in effect, asks us to ignore the obvious wisdom of the Burns court's admonition and to read into the Agreement a representation that nowhere appears in the text. We are unwilling to freelance in this fashion. See id. (warning that the defense, like the prosecution, "must be alert to the need for clear and explicit articulation of all pertinent terms in any plea agreement").
Two other considerations buttress the government's position that the Agreement should be read as written. In the first place, soft-pedaling the integration clause and slipping an antecedent oral promise into the text would render other of the Agreement's relevant passages, such as paragraph 11, entirely nugatory. And this would contravene the rule that plea agreements, like contracts generally, should be construed where possible to give effect to every term and phrase. See Feinberg v. Insurance Co. of N. Am., 260 F.2d 523, 527 (1st Cir. 1958) ("In construing a contract, we must give reasonable effect to all terms whenever possible.").
In the second place, inserting a new promise into the Agreement would turn the change-of-plea colloquy into a farce. At that time, the district court placed the appellant (who was assisted by counsel and does not question their effectiveness) under oath and carefully questioned him. See generally Fed. R. Crim. P. 11(e). The appellant assured the court that the prosecution had made no promises to him apart from those that were written explicitly into the Agreement. Although Alegra's appellate counsel tries to slough this off as a legal fiction and admonishes us that all defendants prevaricate during change-of-plea colloquies, courts cannot operate on the assumption that parties feel free to lie with impunity in response to a judge's interrogation. We believe, therefore, that a defendant who asserts a fact in answer to a judge's question during a change-of-plea proceeding ought to be bound by that answer, absent exceptional circumstances (say, for example, the emergence of newly discovered evidence that places what was reasonably thought to be a fact in a different light). See, e.g., United States v. Doyle, 981 F.2d 591, 594 (1st Cir. 1992); United States v. Butt, 731 F.2d 75, 80 (1st Cir. 1984). We see no reason either to deviate today from this salutary rule or to give the appellant the benefit of the long-odds exception to it.
The appellant attempts in several ways to denigrate the effect of the integration clause and the other circumstances we have mentioned. Citing United States v. Rounsavall, 128 F.3d 665, 668-69 (8th Cir. 1997), he argues that oral representations routinely are used to augment written plea agreements. That case (in which the precise wording of the plea agreement is never discussed) simply does not stand for the proposition that prior oral representations may trump unambiguous language in a plea agreement that expressly purports to be integrated. Moreover, the same court elsewhere has indicated that it will rely on extrinsic evidence only when, after considering a plea agreement as a whole, the parties' intent remains ambiguous. See Kelly, 18 F.3d at 616.
The appellant also cites United States v. Leonard, 50 F.3d 1152 (2d Cir. 1995), for the same proposition. The case makes no such holding. While the plea agreement there apparently contained an integration clause, the only relevant issue before the appellate court concerned a much different question: the good faith vel non of the government's decision not to move for a downward departure.4 See id. at 1157-58. So, too, for obvious reasons, we find inapposite the appellant's citations to a line of cases in which courts have deemed terms outlined in transmittal letters accompanying plea agreements to be part and parcel of those agreements. See, e.g., United States v. Garcia, 956 F.2d 41, 44 (4th Cir. 1992); United States v. Melton, 930 F.2d 1096, 1098-99 (5th Cir. 1991).
When all is said and done, the Agreement's integration clause -- paragraph 22 -- withstands the appellant's bombardment. Accordingly, we hold that it is not reasonable for Alegra to seek the benefit of a prior oral representation by the government after he signed a fully integrated writing that did not contain the claimed representation, and expressly affirmed to the district court in the change-of-plea colloquy that he had not been influenced by extrinsic representations of any kind.
Although the plain language of the Agreement provides no succor and the effort to supplement it fails, the appellant tries an end run. He posits that, in construing plea agreements, courts should imply a duty of good faith in performance. Thus, even though a plea agreement states unambiguously -- as this one does -- that the government retains absolute discretion with respect to the filing of a section 5K1.1 motion, the accused is entitled to expect that the government will honestly evaluate the appropriateness of seeking a downward departure. In his peroration, the appellant asserts that the government thwarted this expectation and that the district court erred by not holding an evidentiary hearing.
This argument is not new. United States v. Garcia, 698 F.2d 31 (1st Cir. 1983) -- not cited to us by either the appellant or the government -- deals effectively with it. That case arose before the sentencing guidelines (and, hence, section 5K1.1) went into effect. It involved a written, fully integrated plea agreement in which the government promised, in its discretion, to make a lenient recommendation at sentencing if the defendant's cooperation were complete and truthful. See id. at 35 n.3. We concluded that allowing the government to retain absolute discretion in these circumstances would "render a significant element of the consideration for appellant's change of plea illusory." Id. at 36. We lent substance to this element by requiring the government "to show a good faith consideration of [the defendant's] cooperation," that is, "to set forth in the record sufficient reasons for its belief that [the defendant] has not cooperated fully and that . . . a recommendation [of probation] is not proper." Id. at 35 (quoting decision below).
To be sure, given the passage of time, the emergence of the federal sentencing guidelines, and the Court's decision in Wade v. United States, 504 U.S. 181 (1992), Garcia is arguably distinguishable. But we think that its central concept -- that the government must perform in good faith the discretionary obligations that it affirmatively undertakes in a plea agreement -- remains good law.5 Of course, this does not mean that courts can add material conditions to plea agreements. See, e.g., Garcia, 698 F.2d at 36 (emphatically eschewing such a course). Nor does it mean that every challenge to the government's good faith necessitates protracted proceedings. The government's burden of showing good faith is only a burden of production, not of persuasion. As long as the government satisfies this modest burden, the trial court need go no further unless the defendant makes a substantial threshold showing that the government acted in bad faith. See Kelly, 18 F.3d at 618; United States v. Khan, 920 F.2d 1100, 1106 (2d Cir. 1990); cf. United States v. Catalucci, 36 F.3d 151, 154 (1st Cir. 1994).
A myriad of practical, commonsense considerations recommend this approach. We mention five of them. First, for a court to inquire into the adequacy of a defendant's performance under a plea agreement and assess the good faith of the prosecutor's evaluation, it likely will need to delve into sensitive matters -- a course that ineluctably will have a disruptive effect on the prosecutorial function. Second, the quantum of knowledge about ongoing investigations that is necessary to make an informed decision often may be very high and the process of acquiring that knowledge may be very time-consuming. Third, an uncontrolled good faith exception will provide criminal defendants, after the fact, with virtual carte blanche. Many of them, having little to lose, will depict their performance glowingly, inviting district courts to regard the prosecution's contrary statements as pretextual (thus prompting further inquiry). We doubt that a proliferation of such collateral litigation would square either with the Supreme Court's decision in Wade, 504 U.S. at 185-87, 112 S.Ct. 1840,6or with the orderly administration of the criminal justice system. Fourth, recognizing a duty of good faith on the prosecutor's part creates possibilities for opportunism in a manner that threatens to countermand the goals of the criminal law -- and these possibilities multiply as the ground rules for such challenges become more lax. We made this point emphatically in Doe, when we explained that "[d]efendants, asked for information to incriminate others, have good reasons to fear for their safety and, unless the prosecutor holds the whip hand, the defendant may offer up some information and hold back the more vital balance in the hope that the court will find the government 'unreasonable' and infer 'bad faith.'" Doe, 170 F.3d at 225. Thus, diminishing the bite of the whip through overzealous enforcement of the duty of good faith would be counterproductive. Finally, the path that we have mapped out comports with our oft-stated belief that, in criminal cases, evidentiary hearings should be the exception, not the rule:
We have repeatedly stated that, even in the criminal context, a defendant is not entitled as of right to an evidentiary hearing on a pretrial or posttrial motion. Thus, a party seeking an evidentiary hearing must carry a fairly heavy burden of demonstrating a need for special treatment.
United States v. McGill, 11 F.3d 223, 225 (1st Cir. 1993) (citations omitted); accord United States v. Isom, 85 F.3d 831, 838 (1st Cir. 1996).
In this case, the pertinent portion of the plea agreement is the government promise to consider whether the appellant had rendered substantial assistance, and, thus, merited a section 5K1.1 motion. See supra Parts II(A)-(B). This commitment carried with it an obligation to evaluate the appellant's assistance in good faith (although the "sole discretion" language in which the promise was couched informed the nature of the obligation). The government proffered facially adequate reasons for its conclusion that the appellant had failed to achieve the substantial assistance benchmark: the supplied information was "[on] occasions . . . hearsay and on others . . . just too meager," and also included "self-serving rationalizations" (a characterization that the government punctuated with a telling example). This rejoinder satisfied the government's burden of production. Taken at face value, the appellant's counter-proffer showed that he attended two debriefing sessions with FBI agents and that he was responsive and truthful. He described in some detail information that he gave regarding alleged kickbacks received by a certain bank officer, and referred the government agents to a company that had been transferring large sums of money between Puerto Rico and the Dominican Republic under suspicious circumstances.
We do not believe that the district court erred in deeming this proffer insufficient to warrant further proceedings. Although the appellant professes to be sanguine about the value of the information that he furnished, the record contains no indication that any of it was useful to the government. By like token, the appellant wholly fails to explain how this information relates to any ongoing criminal investigation. The lower court therefore lacked any kind of reliable framework within which it could even begin to assess whether this "assistance" helped the government to any degree, let alone whether it proved "substantial."7
In a last gasp, the appellant calumnizes the government's failure to respond to information contained in a supplementary letter that he transmitted. We need not linger over this correspondence. The appellant again fails to show how that information was any more useful than the material cited in his declaration. In addition, we explained in Doe that the government's failure to pursue such information, without more, amounts at most to carelessness and does not suffice to make out a case of bad faith. See Doe, 170 F.3d at 225-26.
To say more on this point would be supererogatory. We review the district court's decision as to whether to go further, that is, whether to convene an evidentiary hearing on the issue of the government's good faith, for abuse of discretion. See David v. United States, 134 F.3d 470, 477 (1st Cir. 1998). Stripped of rhetorical flourishes, the appellant offers a wealth of conclusory assertions, but no persuasive evidence of either substantial assistance or bad faith. On this gossamer record, we cannot conclude that the sentencing court abused its discretion in ruling that the appellant failed to attain the requisite threshold. It follows inexorably that the government did not breach the duty of good faith in performance that due process imposes.
The appellant's final departure-related argument is that, if the absence of a government motion places section 5K1.1 beyond his reach, the district court, given his cooperation, nonetheless should have departed downward under the general departure guideline, USSG §5K2.0. See generally Koon v. United States, 518 U.S. 81, 94-95 (1996) (discussing the circumstances in which departures under §5K2.0 are proper); United States v. Dethlefs, 123 F.3d 39, 44 (1st Cir. 1997) (same). We need not linger over this importuning. The three courts of appeals that have addressed the question since Koon agree that section 5K1.1 occupies the field and that departures for substantial assistance, however labeled, are available only under section 5K1.1. See In re Sealed Case, 181 F.3d 128, 140-42 (D.C. Cir. 1999); United States v. Solis, 169 F.3d 224, 227 (5th Cir. 1999), petition for cert. filed, ___ U.S.L.W. ___ (U.S. June 3, 1999) (No. 98-9623); United States v. Abuhouran, 161 F.3d 206, 213 (3d Cir. 1998), cert. denied, 119 S. Ct. 1479 (1999).
We had left the question open in a pre-Koon case. See United States v. Romolo, 937 F.2d 20, 25 (1st Cir. 1991). We now answer it, adopt the reasoning of our sister circuits, and hold that a defendant's assistance to the prosecutor cannot serve as the basis for a section 5K2.0 departure. By necessary implication, then, the district court did not err in refusing to depart downward based on Alegra's cooperation.
We turn last to the calculations underpinning the appellant's sentence. For economic crimes like bank fraud, amount of loss is a critical component in formulating the guideline sentencing range. See, e.g., United States v. Rostoff, 53 F.3d 398, 407-08 (1st Cir. 1995); United States v. Tardiff, 969 F.2d 1283, 1285 (1st Cir. 1992). The appellant claims that the district court erred in calculating the pecuniary losses caused by his conduct. Because this challenge cannot be divorced from the facts, we sketch the contours of the offenses of conviction. As is the custom in sentencing appeals, we draw our factual insights from the change-of-plea colloquy, the presentence investigation report, and the transcript of the disposition hearing. See, e.g., United States v. Dietz, 950 F.2d 50, 51 (1st Cir. 1991). In this case, moreover, we also have the benefit of our opinion in related litigation. See Sheils Title Co. v. Commonwealth Land Title Ins. Co., 184 F.3d 10 (1st Cir. 1999).
The appellant served as the president of Bankers Finance Mortgage Corporation (BFMC). In the ordinary course of its business, BFMC made residential mortgage loans. Many of these loans were refinancings. In such a refinancing, BFMC's business plan called for it to pay off the borrower's existing first mortgage and make a new loan secured by a new first mortgage. It then bundled groups of these loans and peddled the packages to large financial institutions (most prominently, Citibank), in each case representing that the purchaser would receive the functional equivalent of a first mortgage, viz., an assignment of BFMC's first mortgage. When a loan was sold, BFMC would send the purchaser an assignment and a certificate of title insurance.8 Simultaneously, BFMC would notify the borrower to make future payments directly to the purchaser. The purchaser would then pay the agreed purchase price to BFMC.
The major snag in this scenario was that, in certain instances, BFMC never paid off the original mortgages. Thus, the purchasers held second rather than first mortgages. Moreover, the mortgagors went into default on the original mortgages, notwithstanding their payments to the purchasers. BFMC's chicanery came to light when some of the original lenders initiated foreclosure actions. The purchasing institutions were in many instances left holding an empty bag.
The indictment focused on BFMC's transactions with Citibank, which lost approximately $3,100,000 as a result of the scheme. Citibank managed to recoup some two-thirds of this amount from BFMC and the appellant, reducing its net loss to roughly $1,200,000. This amount was reimbursed by the title insurer. See Sheils, 184 F.3d at 12-13 (describing scheme and recounting details of title insurer's involvement).
Against this mise-en-scene, the district court concluded that the essence of the criminal conduct more closely resembled theft than simple fraud (in that the appellant took value from Citibank without intending to give anything of value in return). See United States v. Orton, 73 F.3d 331, 334 (11th Cir. 1996) (explaining difference between theft and simple fraud); United States v. Smith, 951 F.2d 1164, 1167 (10th Cir. 1991) (similar); United States v. Kopp, 951 F.2d 521, 528-29 (3d Cir. 1991) (similar); see also United States v. Flowers, 55 F.3d 218 (6th Cir. 1995) (declining to treat check-kiting cases like fraudulent loan application cases); United States v. Frydenlund, 990 F.2d 822, 825-26 (5th Cir. 1993) (similar); cf. United States v. Schneider, 930 F.2d 555, 558 (7th Cir. 1991) (distinguishing between fraud in which the fraudfeasor intends to deprive the victim of the entire value of an object and fraud in which the fraudfeasor returns something of value to the victim). Having reached this conclusion, the court ignored the title insurer's payments and fixed the amount of loss attributable to the scheme at $1,200,000. See USSG §2F1.1(b). This, in turn, dictated the guideline sentencing range and influenced the length of the prison sentence that the court imposed.
We review sentencing determinations under a bifurcated standard. Quintessentially legal questions, including determinations as to the meaning and application of particular guidelines, engender de novo review. See United States v. St. Cyr, 977 F.2d 698, 701 (1st Cir. 1992). The sentencing court's factfinding, however, is reviewed deferentially and will be disturbed only if it is shown to be clearly erroneous. See id. We assume here, favorably to the appellant, that the de novo standard of review obtains.
The appellant's principal objection to the district court's loss calculation is that it failed to take into account the fact that Citibank, because it enjoyed the benefit of title insurance, never ran a risk of losing anything on the mortgage transactions. In the appellant's view, the title insurance functioned essentially as pledged assets, see USSG §2F1.1, comment. (n.8(b)); the transactions between BFMC and Citibank thus were equivalent to fraudulent loan transactions, see id.; and, therefore, any amounts recovered by Citibank from the title insurer must be offset in computing the amount of loss.9
Although the parties debate longiloquently the question of whether the appellant's conduct was tantamount to a series of fraudulent loan transactions, we need not force our way into Procrustean taxonomies to resolve the underlying dispute. Cf. United States v. Riley, 143 F.3d 1289, 1291-92 (9th Cir. 1998) (endorsing economic reality approach to sentencing). Even assuming that this scheme is best characterized as involving fraudulent loans, we find no error in the court's decision not to shrink the amount of loss to reflect the receipt of title insurance proceeds. Insurance, unlike pledged assets, does not diminish the impact of the fraud. Rather, insurance simply shifts the loss to another victim (the insurance company), so it is irrelevant in calculating the amount of loss for sentencing purposes. See United States v. Daniels, 148 F.3d 1260, 1262 (11th Cir. 1998) (per curiam).
We need go no further.10 There is no question but that the appellant engaged in willful misconduct. The mere fact that his victim was insured puts him in a worse, not a better, position from the standpoint of the criminal law: he not only committed fraud, but his knowledge that Citibank had title insurance permitted him to gamble with other people's money. It would be perverse to hold that criminals need not account for fraudulent losses because they know that, regardless of their machinations, their principal victim will be made whole by an insurance company. The sentencing guidelines surely do not compel such a conclusion.11
The government denies certain of these facts, including the appellant's claim that the United States Attorney verbally represented that he would be rewarded with a downward departure motion as long as he told the full and complete truth. The government also maintains that the information the appellant provided during debriefing was minimal, self-serving, and of no help in its ongoing criminal investigations. Because we accept arguendo the appellant's version of the historical facts, we eschew the government's suggestion that we remand for better development of the record.
At oral argument, Alegra's appellate counsel repeatedly suggested that Puerto Rico law controls whether and to what extent parol evidence might be acceptable to vary or explain the terms of the Agreement. This suggestion falls wide of the mark. A plea agreement in a federal criminal case must be interpreted in accordance with federal, rather than local, law. See United States v. Herrera, 928 F.2d 769, 773 (6th Cir. 1991) ("Federal plea agreements must be governed by the Constitution and federal law, otherwise identical agreements would be subject to different interpretations depending upon which state rule was being applied.").
The only other passage in the Agreement that touches on section 5K1.1 motions is paragraph 13, which states that "[t]he United States reserves its option to seek any departure from the applicable sentencing guidelines, pursuant to Section 5K of the Sentencing Guidelines and Policy Statements, or Rule 35(b) of the Federal Rules of Criminal Procedure, if in its discretion, the United States determines that such a departure is appropriate." That language confirms the government's retention of its wonted discretion, paragraph 8 notwithstanding, and thus reinforces the plain meaning of paragraph 11.
We deal with this question in Part II(C), infra.
This point is somewhat controversial. Some courts infer a similar duty of good faith. See, e.g., United States v. Isaac, 141 F.3d 477, 482-84 (3d Cir. 1998); United States v. Stockdall, 45 F.3d 1257, 1260 (8th Cir. 1995); United States v. Lee, 989 F.2d 377, 380 (10th Cir. 1993); United States v. Rexach, 896 F.2d 710, 714-15 (2d Cir. 1990). Others refuse to recognize such a duty. See, e.g., United States v. Aderholt, 87 F.3d 740, 742-43 (5th Cir. 1996); United States v. Burrell, 963 F.2d 976, 984-85 (7th Cir. 1992).
In Wade, the Court held that, even in the absence of a plea agreement, "federal . . . courts have authority to review a prosecutor's refusal to file a substantial-assistance motion and to grant a remedy if they find that the refusal was based on an unconstitutional motive." 504 U.S. at 185-86. As here, the appellant sought an order of remand to develop his claim that in light of his extensive cooperation the government acted "arbitrarily" or "in bad faith." Id. at 186. The government conceded that relief would be appropriate if the prosecutor's refusal were not rationally related to any legitimate government end. See id. Nonetheless, the Court denied the appellant's request for remand, holding that evidence of the extent of a defendant's assistance to the government would be per se insufficient to show irrationality: "The Government's decision not to move may have been based not on a failure to acknowledge or appreciate Wade's help, but simply on its rational assessment of the cost and benefit that would flow from moving." Id. at 187. The Court's complete indifference to evidence of Wade's cooperation and its willingness to postulate a benign motive for the government's decision make it clear that the Court had no intention of opening the floodgates to a deluge of sentencing-related minitrials.
The government describes the information as "too little and too far removed from the facts." The appellant suggests that the agents' unfamiliarity with the case explains the government's negative assessment. This is sheer speculation, and falls well short of what is needed for a substantial threshold showing of bad faith.
The appellant procured the title insurance certificates fraudulently: he would forward to the title insurer's agent a copy of a check made payable to the original lienholder, representing that a portion of the borrower's loan proceeds were being used to discharge the debt. However, the appellant would not send the check to the lienholder, but, rather, would conceal it in a desk drawer. See Sheils, 184 F.3d at 13 n.4.
Application Note 8(b), formerly Application Note 7(b), provides in pertinent part:
In fraudulent loan application cases and contract procurement cases, the loss is the actual loss to the victim (or if the loss has not yet come about, the expected loss). For example, if a defendant fraudulently obtains a loan by misrepresenting the value of his assets, the loss is the amount of the loan not repaid at the time the offense is discovered, reduced by the amount the lending institution has recovered (or can expect to recover) from any assets pledged to secure the loan. . . .
* * *
We note that the district court may well have underestimated the amount of loss by deducting the payments made to Citibank by the appellant and BFMC after the fraud was discovered. See United States v. Scott, 74 F.3d 107, 111-12 (6th Cir. 1996); United States v. Shaffer, 35 F.3d 110, 114 (3d Cir. 1994). Because the government does not pursue this point, we do not decide it here.
The appellant also contends that the district court erred in failing to depart downward on the ground that the discerned amount of loss ($1,200,000) overstated the seriousness of the offense. This contention borders on the frivolous, and we reject it out of hand.