13 F.3d 489
UNITED STATES of America, Appellee,
Aaron STERN, Defendant, Appellant.
UNITED STATES of America, Appellee,
Lawrence GORDON, Defendant, Appellant.
Nos. 92-2300, 93-1047.
United States Court of Appeals,
Heard Sept. 9, 1993.
Decided Jan. 20, 1994.
Martin D. Boudreau, North Quincy, MA, for appellant Aaron Stern.
Lawrence Gordon, on brief pro se.
Paul G. Levenson, Asst. U.S. Atty., with whom A. John Pappalardo, U.S. Atty., Boston, MA, was on brief, for U.S.
Before BOUDIN, Circuit Judge, COFFIN and CAMPBELL, Senior Circuit Judges.
BOUDIN, Circuit Judge.
The Miller Act, 40 U.S.C. Secs. 270a-d requires all contractors bidding for government construction contracts in excess of $25,000 to post performance and payment bonds, and the Air Force further requires that a bid bond accompany the bid itself.1 In order to qualify for consideration, contractors must submit bonds issued by companies approved by the United States Treasury and listed in Treasury Department Circular 570, commonly called the "T-list." See 48 C.F.R. Sec. 28.202 (a)(1). The bonds must also be submitted on standard government forms: SF 24 (bid bond), SF 25 (payment bond) and SF 25A (performance bond).
Defendant Lawrence Gordon was the head of Tower Associates, Inc., a Winchester, Massachusetts, construction company seeking to secure a contract to renovate a photography laboratory at Hanscom Air Force Base in Bedford, Massachusetts. In September 1988 Tower submitted the low bid for the project, offering to perform the renovations for $1,000,200. This bid was accompanied by a bid bond issued by Continental Surety Company, a surety or purported surety that did not appear on the T-list and which apparently had no assets. The Air Force employee responsible for overseeing the bidding process, Lorraine McLoughlin, did not at first notice this problem and Tower was awarded the contract on September 30, 1988.
When shortly thereafter McLoughlin learned that Continental was not an approved issuer, she called Gordon and informed him that Tower's payment and performance bonds would have to be written by a T-listed company. On October 17, 1988, Gordon presented a payment and performance bond purportedly issued by Amwest Surety Insurance Co., a company that did appear on the T-list. The bond bore Tower's seal, as well as the signatures of Gordon and one "Alan Stime," who was listed as Amwest's attorney-in-fact. The bond was accompanied by a power of attorney, purportedly from Amwest, also signed by "Alan Stime."
The Amwest bond and power of attorney were counterfeits fabricated by James Grier, the principal of Continental. Grier later testified at trial that he produced the bogus documents at Gordon's request. Sandra Catalano, a Tower employee, testified at trial that she was present when the bond was signed by Gordon and defendant Aaron Stern, who signed the bond as "Alan Stime." At trial, an Amwest official testified that the bond was not a genuine Amwest bond and that no "Alan Stime" was or ever had been an authorized attorney in fact for Amwest.
The Air Force rejected the phony bond after McLoughlin noted that some of the signatures on the bond appeared to be facsimiles and that the purported Amwest seal was poorly impressed and illegible. On October 19, 1988, McLoughlin requested that Tower resubmit its bonds and enclosed standard government bond forms. The Air Force received a second set of bonds, on the government forms, from Tower on October 24, 1988. These bonds were also purportedly issued by Amwest, but the typed name of the attorney-in-fact under the "Alan Stime" signature was "Aaron Stern." By this time, McLoughlin had been told by an Amwest employee that Amwest "had never heard of Tower Associates."
Rather than accept the bonds, McLoughlin forwarded them to the Air Force Office of Investigations and sent Tower a notice to cure. On November 18, 1988, McLoughlin notified Gordon of her communications with Amwest. After requesting an extension of time to submit new bonds, Tower sent McLoughlin a third set of bonds on December 13, 1988, explaining that Tower "[had been] given a bond which proved invalid." This third set of bonds, like the original bid bond, was issued by Continental and signed by Aaron Stern as attorney-in-fact. As Continental was still not on the T-list, McLoughlin rejected the bonds.
The Air Force terminated Tower's award on March 3, 1989, and eventually awarded the contract (without rebidding) to Fellsway, Inc., which had submitted the second lowest bid. On June 13, 1991, Gordon and Stern were charged in a multi-count indictment with the following offenses:
Count 1 Gordon and Stern were both charged with conspiring to defraud the United States in the solicitation and award of the construction contract at Hanscom Air Force Base. 18 U.S.C. Sec. 371 (conspiracy to defraud).
Count 2 Both defendants were charged with counterfeiting the October 17, 1988, payment and performance bond purportedly issued by Amwest Surety Insurance Company. 18 U.S.C. Sec. 494 (making, uttering or presenting counterfeit bond).
Count 3 Gordon was charged with knowingly presenting the same counterfeit bond to the Air Force. 18 U.S.C. Sec. 494.
Count 4 Gordon and Stern were both charged with uttering to the Air Force a counterfeit power of attorney. 18 U.S.C. Sec. 495 (making, uttering or presenting counterfeit power of attorney).
Count 5 Both defendants were charged with false statements in completing and submitting Standard Form 25, the government form for performance bonds. 18 U.S.C. Sec. 1001 (false statement statute).
Count 6 Both defendants were charged with false statements in completing and submitting Standard Form 25A, the government form for payment bonds. 18 U.S.C. Sec. 1001.
After a jury trial, Stern was convicted on counts 1 and 4, and acquitted on count 2. Gordon was convicted on count 3, and acquitted on counts 1, 2 and 4. Both defendants were convicted on Counts 5 and 6. Gordon moved for a new trial on June 19, 1992, arguing that the verdict was internally inconsistent and that the government had withheld material exculpatory evidence. The district court denied this motion on December 18, 1992.
On October 13, 1992, the district court sentenced Stern to a 60-day term of imprisonment, along with a period of supervised release. Gordon was sentenced on November 24, 1992, to a 30-day term of imprisonment and nine months of home confinement. Both defendants were also held jointly and severally liable for restitution.2 These consolidated appeals followed. Stern's counsel has briefed and argued the case; Gordon, with this court's permission, has relied upon his district court filings in support of a new trial.
In this court both defendants claim that the jury verdicts are internally inconsistent. Gordon argued in the district court that since he was acquitted (under count 2) of counterfeiting the October 17 bond, it was inconsistent for the jury then to convict him (under count 3) of uttering the same counterfeit bond. Stern points to his own acquittal of the charge of counterfeiting the bond (again under count 2) and protests that this acquittal undermines the jury verdict convicting Stern (under count 4) of uttering a forged power of attorney in support of the same bond.
Both objections lack merit. As the defendants purport to recognize, general jury verdicts may not normally be set aside for inconsistency as between counts. United States v. Powell, 469 U.S. 57, 64-65, 105 S.Ct. 471, 476, 83 L.Ed.2d 461 (1984); United States v. Bucuvalas, 909 F.2d 593, 597 (1st Cir.1990). The reasons are explained by Judge Friendly in United States v. Maybury, 274 F.2d 899 (2d Cir.1960), and do not need repeating. Maybury is relied on by both defendants because the appeals court there did set aside verdicts as inconsistent. But the inconsistent verdicts were there rendered by a judge in a jury-waived trial. The whole point of Maybury is that (for both practical and historical reasons) the general verdict by a jury is a special case but a requirement of consistency does apply to written findings made by a single judge. Id. at 903.
We need not discuss other inconsistent-verdict cases cited by defendants because, as it happens, there is no necessary inconsistency in the verdicts in this case. As to Gordon, there was evidence from Grier that he (Grier) fabricated the October 17, 1988, bond, but also evidence that Gordon knew it was counterfeit and nevertheless presented it to the Air Force. The jury may have supposed (quite wrongly) that Grier's admission entirely lifted responsibility for the counterfeiting from Gordon's shoulders but also permissibly believed that Gordon uttered the bond knowing it to be forged.3
Similarly, the acquittal of Stern on the charge of counterfeiting the same bond, again quite possibly because of Grier's admission, is in no way inconsistent with Stern's conviction for uttering to the Air Force the companion power of attorney document knowing it to be forged. In addition to other evidence to support the uttering conviction, there was direct testimony from Sandra Catalano that Stern himself signed the bond using the phony signature "Alan Stime."
Whether or not Stern, Gordon or both could have been convicted of procuring or participating in the counterfeiting of the bond itself is irrelevant. It is sufficient to dispel the taint of inconsistency that a rational jury could easily have acquitted on count 2 while convicting Gordon on count 3 and Stern on count 4. And, as explained at the outset, inconsistency would not in any event undermine the convictions so long as they themselves were, as they are here, supported by sufficient evidence. See Powell, 469 U.S. at 67, 105 S.Ct. at 478.
Next, Gordon claimed in the district court that the government violated its obligation, under Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), to produce exculpatory evidence by failing to turn over the grand jury testimony of Grier, whom Stern called as a witness. The background is this: Grier, called by Stern as a defense witness, was then cross-examined by Gordon's counsel--not by the government. On this cross-examination, Grier proceeded to testify that Gordon had asked Grier to forge the October 17, 1988 bond. Although Gordon was acquitted of that forgery, he was convicted of uttering the forged bond, and the Grier testimony--elicited by Gordon's own counsel--may have helped to confirm Gordon's knowledge of the forgery.
After conviction Gordon learned that Grier, testifying in the grand jury prior to trial, had himself denied any knowledge of the bond. Gordon moved for a new trial, contending that the government's failure to produce the grand jury testimony at trial as impeaching evidence violated its obligation under Brady. In pre-trial requests Gordon had asked in general terms for Brady material, but he had never specifically requested production of prior statements or testimony by Grier. The district court denied the new trial motion, citing United States v. Pandozzi, 878 F.2d 1526 (1st Cir.1989), and United States v. Carrasquillo-Plaza, 873 F.2d 10 (1st Cir.1989).
Under the Jencks Act, 18 U.S.C. Sec. 3500, the government is required to produce prior statements by its own witnesses, whether or not the statements are exculpatory. And, if a statement is itself exculpatory, the government under Brady is normally required to produce it, regardless of whether it is made by a trial witness. See Brady, 373 U.S. at 86, 83 S.Ct. at 1196. Here, says the government, Grier's grand jury testimony was not a prior statement by a government witness, nor was it exculpatory in the sense that it disproved Gordon's guilt.
Thus, in the government's view, the grand jury testimony is merely newly discovered impeaching evidence. Under the Wright-Martin standard for a new trial based on newly discovered evidence, the ordinary requisites for a new trial on this ground are specific and demanding.4 Gordon may not have met any of the requisites, and he certainly did not meet the requirement that the newly discovered evidence be (at least in the normal case) "not merely ... impeaching." Martin, 815 F.2d at 824 (quoting Wright, 625 F.2d at 1019).
Thus Gordon's only hope is to argue that the grand jury testimony did have to be produced under Brady in which event, as the government notes, the Wright-Martin standard does not invariably apply. See United States v. Sanchez, 917 F.2d 607, 617 (1st Cir.1990), cert. denied, 499 U.S. 977, 111 S.Ct. 1625, 113 L.Ed.2d 722 (1991). Rather, there is a general obligation to produce exculpatory testimony and failures to comply are reviewed after the fact under the standard of United States v. Bagley, 473 U.S. 667, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985). Bagley is somewhat opaque, there being no single majority opinion, but reversal may be warranted where there is a "reasonable probability" that the undisclosed evidence would have altered the outcome. Id. at 682, 105 S.Ct. at 3383 (plurality opinion).
Here the grand jury testimony was not exculpatory in the sense that, reading the bare language before trial, an assistant U.S. attorney would think it helpful to Gordon. After all, in the testimony Grier denied knowledge of the forgery, a position that could be neutral in its impact on Gordon or even potentially harmful to Gordon (in shifting responsibility to someone other than Grier). Only after Grier, as a defense witness for Stern, named Gordon as the procurer of the forgery did the grand jury testimony take on a potential as impeaching evidence.
The government protests that it has no ongoing obligation to monitor the testimony of defense witnesses and to seek out prior statements in its files that may in the course of trial turn out to have impeachment value when defense testimony injures a defendant. Even if it did have such an obligation in some instances, there would be serious questions--not easily answered on this record--whether in this case Gordon's own access to Grier negated the obligation, see United States v. Hicks, 848 F.2d 1, 3-4 (1st Cir.1988), and whether the failure of Gordon to request the grand jury testimony is also fatal to the Brady claim. Cf. 18 U.S.C. Sec. 3500(b) (request by defendant required under Jencks Act); Carrasquillo-Plaza, 873 F.2d at 13 (failure to make a specific request for alibi witness statements).
We think that these interesting questions had best await another occasion. It is enough in this case that the impeaching evidence, even if made available to Gordon, could not conceivably have altered the outcome. See generally Pandozzi, 878 F.2d at 1528-30. The jury acquitted Gordon on the counterfeiting count despite Grier's direct inculpation of Gordon; and the knowledge element of the uttering count, on which Gordon was convicted, was confirmed by the direct testimony of another witness, Catalano, as well as much else in Gordon's conduct. Under Bagley, the impeaching evidence does not remotely "undermine confidence" in the outcome. 473 U.S. at 682.
The final issue in this case is the most perplexing and relates solely to sentencing. Under the Sentencing Guidelines,5 the amount of "the loss" is a specific offense characteristic of crimes involving fraud or deceit, and the base offense level (6 levels) is increased by a specified number of levels (from 0 to 11 additional levels) depending on the amount of the loss. U.S.S.G. Sec. 2F1.1 (1988). The adjusted offense level, together with criminal history, determines the sentencing range, and actual loss may also be the basis for a restitution order. 18 U.S.C. Sec. 3663(b)(1).
In this case, following the pre-sentence report, the district court found that the loss to the Air Force of the fraudulent activities in this case was "the difference between the bid price [by Tower] and the award [to Fellsway, the second lowest bidder] for $88,477, increased by the restated administrative costs of $250" involved in reawarding the contract to Fellsway. The resulting figure, $88,727, increased the base offense level from 6 to 11, U.S.S.G. Sec. 2F1.1(b)(1) (1988), and this in turn was increased to 13 by the addition of two more levels for "more than minimal planning." Id. Sec. 2F1.1(b)(2).
The loss attributed to Stern may well have had no effect on his sentence of confinement; a downward departure, based on assistance to the government, reduced his confinement to below the minimum range that would have prevailed if no loss had been attributed to him. Gordon's sentence of confinement, also based on a downward departure, might or might not have been affected by a lower loss figure, depending on how small a loss was imputed.6 But based on the imputed loss of $88,727, both defendants were ordered to pay restitution--Stern to pay $88,727 and Gordon $80,000--so the importance of the loss figure is obvious.
At this point, some procedural history is needed. In the district court, both defendants challenged the $88,727 loss figure on somewhat different grounds. Then, while these appeals were pending, someone apparently happened upon Judge Posner's decision in United States v. Schneider, 930 F.2d 555 (7th Cir.1991), which undoubtedly made the government uneasy about the loss calculation in this case. In any event, the government and defendants entered into a joint stipulation, proposing that this court remand the case, before deciding the merits, for resentencing in light of Schneider; and the parties stipulated further that
1. The parties jointly stipulate that the best readily calculable measure of the loss from the offenses of conviction, for purposes of applying the Sentencing Guidelines, is $20,450. This amount includes both an attempted gain of $20,200 (the purchase price for a genuine bond of the kind that was forged in this case), and actual consequential losses of $250 (the immediately identifiable administrative costs associated with the re-awarding the Photolab contract). For purposes of ordering restitution, only the actual loss figure, $250 is subject to restitution.
2. The United States further reserves the right to argue that the stipulated amount--while representing the best readily ascertainable estimate of loss--understates the full extent of the loss in question. The United States further reserves the right to argue that, given the small sum of restitution payable, fines should be imposed upon each defendant, in an amount to be fixed in light of the defendants' resources.
This court denied the motion to remand, believing that the challenges to the convictions ought to be decided before any remand for fine-tuning the sentences. Stern, taking the view that the stipulation was binding only if this court ordered an immediate remand, has argued in his brief in this court that no loss, apart perhaps from the $250 involved in shifting the award to the second bidder, has been shown by the government. The government adheres to its request for a remand in accordance with the stipulation, arguing that Stern has waived the contention he now makes. Gordon has urged nothing beyond the stipulation.
It would be a hazardous venture to lay down abstract rules as to how "loss" is to be calculated under the governing guideline even if the focus were narrowed to false statements in bid documents. As Judge Posner made clear in Schneider, the underlying facts of individual cases may differ widely, and even in comparable situations, what proof is available as to specific items of loss will vary from case to case. See 930 F.2d at 557-59. Further, we note that the deceptively simple notion of "loss" is elaborated under the guideline to include situations of foreseeable or intended losses, see U.S.S.G. Sec. 2F1.1, application note 7 (1988), and in later versions to cover various specific types of fraud, e.g., id., application note 7(e) (1993) (Davis-Bacon Act fraud).
If we agreed with the district court's original calculations of loss, we would not remand the case regardless of the stipulation of the parties, since the parties cannot by agreement create error where none exists. On this record, however, we think that there is no basis for mechanically measuring the loss as the difference between the two bids. The problem is that the government never showed that it could have secured a bid from a properly bonded contractor at the price offered by Tower. Without such evidence, it is hard to see how the government can measure its loss on the theory that, but for the fraud, it would have enjoyed that initial low price.7
On certain facts--say, a general increase in the level of second round bids after a rebidding due to fraud--a calculation of loss based on the differential between a tainted first-round best bid and a higher second-round best bid might be entirely persuasive. But here the government simply took the second best first-round bid with no rebidding; and its administrative costs in shifting the award from Tower to Fellsway were admittedly minimal ($250). The probation reports and the government urged the loss figure adopted by the district judge, and the defendants while protesting did little to undermine it. Yet the government itself no longer supports the $88,275 figure, and we can find no basis to sustain it on the present record.
Conversely, we reject Stern's "no loss at all" argument, at least so far as concerns the guideline calculation. The section 2F1.1 guideline commentary, from the 1988 version (application note 7) to the present 1993 version (id.), has included the "probable" or "intended" or "expected" loss threatened by a defendant's conduct as an alternative measure of loss if that figure can be determined and is larger than actual loss. The evident purpose of the alternative is to be certain that attempted fraud does not escape all adjustments based on magnitude merely because the fraud miscarried. We think that this is just such a case where the foreseeable loss exceeded the actual loss.
It was plainly foreseeable at the time of the fraud that the Air Force might be deceived by the phony Amwest bond and might finally award the contract to Tower. At that point, it is hard to know the precise risk of loss imposed on the government, for it would depend on many circumstances including the likelihood that Tower would flawlessly complete the contract. Yet "the amount of loss need not be precise," U.S.S.G. Sec. 2F1.1, application note 7 (1988); and, broadly speaking, to inflict a phony construction-contract bond on the government exposes the government to the average cost of failure to perform the contract adjusted by the average risk of failure to perform. That adjusted figure should also be the approximate price of the average bond for such a project. Of course, the pertinent figure could be higher if the government sought to prove that Tower was a high-risk contractor and would have been charged more for a valid bond.
To assume that the phony bond might well be accepted is, in a case like this one, entirely fair. It is the defendants' intended outcome and, given possible carelessness by administrators or merely a good forgery, it is normally a realistic likelihood. From the standpoint of the guideline's "intended or probable loss" criterion, we see nothing wrong with the use of the cost of a valid bond as a fair proxy for the potential loss caused by the uttering of the phony bond and forged power of attorney. In this case, the potential was not realized, but it was still intended or reasonably likely and thus a proper measure of loss under the guideline.
Consequently, we think that the principle underlying the first paragraph of the quoted stipulation is a legitimate basis in this case for calculating loss under the guideline; whether the $20,200 figure is binding on Stern is a matter that can be resolved on remand if Stern seeks to disclaim the stipulation.8 8] Of course, the government did not offer such evidence of bond cost in the original sentencing proceeding; but where a sentence is vacated and remanded for redetermination under correct principles, the government is not automatically foreclosed from offering evidence pertinent to the newly announced rule. See United States v. Sepulveda, 15 F.3d 1161, 1199 n. 31 (1st Cir. 1993).
Restitution is a different matter. We agree with the government's concession that the intended or probable loss cannot be the measure of restitution: it is one thing to base a criminal sentence on the magnitude of threatened harm but quite another to "restore" to the government money that it never lost. Here, the actual loss is only the $250 administrative cost of reawarding the contract. Given the common interrelationship between fines and restitution, we see no reason why the government should not be allowed to argue for a fine on remand.
Accordingly, the judgment of conviction in each case is affirmed, and the sentence and the orders of restitution in each case are vacated and the cases are remanded for resentencing in accordance with this opinion. Further proceedings on remand are for the district court to determine in the first instance.
It is so ordered.
"Bid bonds" ensure that a contractor will in fact undertake the contract if its bid is accepted; "performance bonds" guarantee that the contractor will complete the project in accordance with the specifications; and "payment bonds" ensure that those who furnish labor and materials for the project will be paid
Stern's sentence was stayed pending appeal. The district court denied Gordon's motion for a similar stay on January 19, 1993, but this court stayed the payment of Gordon's restitution pending appeal on August 24, 1993
Gordon's claim that the conspiracy acquittal (under count 1) is inconsistent with the uttering conviction (under count 3) is even more far-fetched. True, the uttering was charged as an act in furtherance of the conspiracy. But a literal minded jury might believe that the uttering itself was proved amply but that the "agreement" element of conspiracy had not been established
A new trial will ordinarily be denied absent a showing that
(1) the evidence was unknown or unavailable to the defendant at the time of the trial; (2) failure to learn of the evidence was not due to lack of diligence by the defendant; (3) the evidence is material, and not merely cumulative or impeaching; and (4) it will probably result in an acquittal upon retrial of the defendant.
The pre-sentence report referred to the 1988 guidelines, which were in effect when the crime was committed. Current practice would normally invoke the guidelines in effect at the time of sentencing--the 1991 version for Stern and the 1992 version for Gordon--barring any ex post facto problems. See Isabel v. United States, 980 F.2d 60, 62 (1st Cir.1992). But since the 1991 and 1992 guidelines employ a new loss/increase-in-level table which would have resulted in a higher base offense level for both Stern and Gordon than provided for under the 1988 guidelines, we cite to the 1988 version. See United States v. Harotunian, 920 F.2d 1040, 1042 (1st Cir.1990)
Gordon's departure was based on the theory that the Air Force could have rebid the entire contract for $10,000 (a figure supplied by the Air Force), instead of merely awarding it to the previous second lowest bidder. Cf. United States v. Gregorio, 956 F.2d 341, 344-48 (1st Cir.1992). The government did not appeal this departure
It is true that if the Amwest bond had been a real bond instead of a forgery, then the government would have enjoyed the original low price; and in this sense the forgery may seem like the cause of that loss. But there is no evidence that Tower Associates could ever have secured a real bond by a T-list surety or, if it could, that its bid would have been as low. Indeed, Continental seems to have been an off-shore "front" for contractors who could not get T-list bonding
The district court is not required to accept a stipulation on an issue of fact pertinent to sentencing, but is free to do so given the apparent reasonableness of the proposed figure. This assumes, of course, either that Stern withdraws his disclaimer (as he would plainly be wise to do) or that the district court decides that his initial acceptance of the stipulation is binding in any event