707 F.3d 68 | 1st Cir. | 2013
THOMPSON, Circuit Judge .
Overview Abdela Tum asks us to overturn his bench-trial convictions for violating and conspiring to violate the federal wire-fraud statute, a law that (at the risk of oversimplification) criminalizes a scheme to defraud involving an interstate-wire communication. See 18 U.S.C. §§ 1343 and 1349. Tum's attack basically comes in two waves. Convinced that the evidence failed to show that he had caused an interstate-wire transmission, he argues first that his wire-fraud convictions cannot stand. Reminding us that conspiracy involves a knowing agreement between two or more parties to violate some other law, he protests that the evidence did not show the existence of a coconspirator, so the conspiracy conviction must fall too.
Winning a sufficiency challenge like this is no easy thing, because we must examine the facts and inferences in the light most favorable to the government, even though our review is de novo ("de novo" being another way of saying that we give the judge's legal ruling a fresh look). See, e.g., United States v. Guerrier, 669 F.3d 1, 7 (1st Cir. 2011). Tum clearly faces a steep uphill fight. And it is one that he must lose, for reasons that will shortly become clear.
Hiding the Truth Tum worked for Barber Foods in Maine as a production worker for almost 16 years, from 1990 to 2006. His wife, Sherifa Hussen, worked there too. Asthma left him unable to do his job. And eventually he started receiving short- and then long-term disability payments through a benefits policy issued to Barber Foods by Unum Life Insurance Company of America, which, the parties tell us, is a Maine-based firm. When Tum started working for Barber Foods, he spoke very little English. But when he was looking to jump-start the disability-payment process about a decade and a half later (we do not know exactly how, but at some point he did learn about the benefits policy, obviously), he had no trouble communicating with Barber Foods's benefits manager.
For a person in Tum's then-position, the policy in play here does a number of important things: It considers him "disabled" (in insurance-speak) if he suffers an earning loss of "20% or more" caused by a "sickness or injury" that restricts him "from performing the material and substantial duties" of his "regular occupation." It also lets him work while collecting disability benefits but warns that Unum will reduce or stop payments depending on how much he earns (the particulars of this are unimportant). On this last point the policy stresses that Unum "may require" him to prove his earnings. Over the next few months Unum would remind Tum more than once that he had to remain disabled as defined by the policy to keep getting benefits.
Skipping over details not relevant to the issues on appeal, we see that Unum began sending Tum checks monthly sometime in 2006 – checks that Tum quickly deposited into an account at KeyBank in Maine. In June 2007, however, Tum asked Unum if it could deposit the funds directly into his KeyBank account instead. Unum said yes, triggering a series of events that took place in various states, which we thumbnail this way. Unum inputted Tum's bank account and routing numbers into a "batch" system – i.e., a system where information is collected, stored, and processed later, rather than in real time. See Harry Newton, Newton's Telecom Dictionary 180 (26th ed. 2011). The batch system's server is in South Carolina. JP Morgan in Florida then verified Tum's data as part of the next phase in the review process. Everything checked out just fine, and so in July 2007 Unum started electronically transferring funds it had with JP Morgan in Illinois to a KeyBank department in Ohio – funds that passed through the Federal Reserve system (there is no Federal Reserve branch in Maine, apparently) – but only after JP Morgan in Florida gave its Illinois counterpart the go-ahead each time. And that is how things pretty much went until December 2009, when the FBI got involved. But we have gotten ahead of our story and so must back up a bit.
Tum had a secret, and it was a doozy: He had been working for Home Health Care Solutions ("HHCS," for short) since March 2007 as a driver. A slew of documents – e.g., independent- contractor agreements between HHCS and Tum, invoices from Tum to HHCS, copies of checks from HHCS to Tum, a letter from HHCS recording payment advances to Tum, documents from HHCS to Tum's mortgage provider verifying his employment – showed that, and this too: that, given his HHCS earnings, Tum was raking in disability benefits that he was not entitled to (as the government argues and Tum does not contest).
Trying to do his best to keep from being found out (an obvious inference from the record), Tum did tell an Unum disability specialist in December 2007 that he was thinking about getting a part-time job. But he said not a word about the HHCS job that he already had. Anyway, the specialist told Tum that if he did return to work, he needed to call back with details "regarding his employer, his new position and his pay" so Unum could verify his disability status. Understood, Tum said (or something to that effect).
Hard on the heels of this exchange, a number of significant things happened: Tum stopped submitting invoices to HHCS – the next invoice bore his wife's name. HHCS created dual sets of invoices for work completed before the December 2007 conversation, one in Tum's name and the other in Hussen's, and started cutting checks payable to Hussen, not Tum. HHCS generated an independent-contractor agreement in Hussen's name too. Also, an HHCS accountant prepared a 1009 form for Hussen – but not for Tum – that listed her supposed 2007 earnings. And Tum had a tax specialist use that document in preparing the couple's joint return, which the specialist then e-filed to the IRS. Adding to the intrigue, Hussen would later tell federal agents that she never worked for HHCS – a story the judge would later credit.
Exchanges like the one Tum had with Unum in December 2007 happened a lot over the next two years. Here are some highlights (or lowlights, if you will):
In January 2008 Tum told another Unum employee that he was delivering newspapers for the Portland Press Herald, pocketing 15 cents per delivery, and he promised to provide Unum with his Portland Press Herald paystubs. But he again said nothing about his HHCS work – even though by this time he had been with HHCS for nine months.
Also, Unum sent Tum several letters in 2008 asking him to document his monthly earnings. Having not gotten a proper and timely response, Unum stopped paying disability benefits to him in October 2008. At last stirred to action, Tum faxed a handwritten note to Unum in December 2008, explaining that his "income is very low" because he has not been "able to do meaningful work." He "was delivering papers" for the Portland Press Herald, he added, making "$350 biweekly." Enclosing an "income report," he said that that number did not really reflect his true earnings, because he had to pay his work-related expenses (gassing up his car to make deliveries, for example) out of his own pocket. Not a peep about HHCS, though.
About two weeks later, Unum sent Tum a letter saying that it had "reopened your claim and released [your] benefits . . . ." Tum wrote back in January 2009, thanking Unum for "being concerned about people" and for "helping us to stay in our home." Discussing his income, Tum stressed that he presently earned "under $10,000 a year." "I have applied everywhere for employment," he continued, but "[s]o far there has been no response. Right now I am trying to get a license to become a taxi cab driver. I believe this will make me self sufficient." Following a familiar pattern, Tum's letter gave no hint that he was working with HHCS.
Sometime in April 2009, Unum reviewed Tum's 2008 earnings with a critical eye and found that what he had told them and what he had told the IRS were two different things – he had given Unum a smaller number, a discrepancy, Unum concluded, that resulted in his getting a few thousand dollars more in benefits than he should have. Obviously Unum did not realize how large the discrepancy really was, given how Tum was mum about his HHCS income. Tum did say, though, that he could not pay any money back, as Unum demanded, because the Portland Herald Press had fired him over some late deliveries, adding that his only source of income now was his Unum disability checks and his wife's earnings at Barber Foods – an untruth for sure, given the HHCS income that he was still keeping from Unum.
Arrest and Fallout Tum's scheme came to a screeching halt in 2010, when federal authorities filed a criminal complaint against him and his wife, Hussen. A federal grand jury later indicted them each on sixteen counts of wire fraud and one count of conspiring with unnamed others to commit wire fraud. See 18 U.S.C. §§ 1343 and 1349. The two waived their right to a jury and asked for a bench [1] trial. The judge granted their request. At the close of the government's case, the pair moved for acquittal on all counts based on insufficient evidence. The judge reserved decision. See Fed. R. Crim. P. 29(b). The duo then presented some evidence in their own defense, calling a couple of witnesses who mostly talked about Tum's limited English language skills, and they renewed their motions for acquittal after they rested. The judge reserved ruling on those motions too, saying he wanted to hear closing arguments first.
Eventually the judge acquitted Hussen of all charges. For starters, he found no evidence that she knew what Unum wanted Tum to do or that his HHCS earnings could affect his benefits. Sure, she probably should have suspected that Tum was trying to scam Unum, the judge added, "but that's not enough." Proof beyond a reasonable doubt is required, he reminded the lawyers, and the evidence against Hussen did not satisfy that standard on any of the counts.
But there was more than enough to find Tum guilty of "schem[ing] to defraud Unum" of "its money" in violation of the wire-fraud act, the judge ruled. And, zeroing in on the HHCS- related documents (contracts, invoices, checks, letters, tax documents, etc.), the judge also found sufficient evidence of Tum's conspiring with HHCS to commit wire fraud. So he rejected Tum's acquittal pleas and found him guilty on all counts.
Which brings us to today, with Tum staking everything on convincing us that the judge erred on the sufficiency issues. We tackle his contentions in the next section, adding a few more details to this saga when needed.
Issues and Rulings First up is Tum's attack on the sufficiency of the evidence behind his wire-fraud convictions – a challenge we review de novo and in the light most favorable to the judge's guilty finding. See, e.g., Guerrier, 669 F.3d at 7.
To prove wire fraud, the government had to show beyond a reasonable doubt his knowing and willful participation in a scheme to defraud and the use of interstate wires to further that scheme. See, e.g., United States v. Denson, 689 F.3d 21, 24 (1st Cir. 2012). The scheme had to "employ material falsehoods" too – i.e., false or omitted statements that a reasonable person would consider important in deciding what to do. Neder v. United States, 527 U.S. 1, 16, 20 (1999). Also, the government need not have shown that he "personally use[d] the wires," but only that "such use was a reasonably foreseeable part of the scheme in which [he] participated." United States v. Woodward, 149 F.3d 46, 63 (1st Cir. 1998) (quoting United States v. Sawyer, 85 F.3d 713, 723 n.6 (1st Cir. 1996), which in turn was quoting United States v. Boots, 80 F.3d 580, 585 n.8 (1st Cir. 1996)) (internal quotation marks omitted); see also United States v. Fermín Castillo, 829 F.2d 1194, 1198-99 (1st Cir. 1987) (explaining that "[w]hether or not the appellant had foreknowledge of the precise series of [electronic communications] . . . is beside the point," and stressing that "[a]s long as some use of [wires] in the course of the endeavor was reasonably to be anticipated, the causation requirement is met").
Taking aim at the interstate-wire requirement, Tum claims that the government's proof here fell short. His thesis goes [2] something like this: Unum has an office in Maine. KeyBank is a Maine bank. And Tum lived in Maine. Consequently, the transferring of benefits had to have been done by intrastate – not interstate – wire communications.
Not so. Recall what happened when Tum asked Unum to send his disability payments via direct deposit. Yes, an Unum employee in Maine keyed Tum's banking information into the batch system. But, devastating to his Maine-centric take on events, the system's server is in South Carolina, and the data was transmitted to Florida for processing/verification. Also, and equally hurtful to his cause, the electronic transferring of funds involved banks in Illinois and Ohio too. In other words, looking at the facts and inferences in the light most helpful to the government, see, e.g., Guerrier, 669 F.3d at 7, we believe that a factfinder could rationally find the use of interstate wires here.
Still hoping against hope to avoid all this, Tum tries to make two points. Neither persuades. His opening argument has three steps. Step 1: KeyBank told its customers in May and June 2007 that going forward, its checking statements would provide more details concerning the wire- transfer process – e.g., the new and improved "wire descriptions will include the wire transfer Sender's Name . . . ." Step 2: The Unum direct deposits lacked that information. Step 3: Ergo, one can infer from this that no interstate-wire transfers happened. What hurts Tum, however, is the standard of review, which compels us to draw all reasonable inferences in the government's favor, not his. See, e.g., Guerrier, 669 F.3d at 8. And the judge had enough evidence to infer – consistent with the government's theory – that an interstate-wire communication actually occurred. That Tum's interstate thesis may be plausible changes nothing, "because the issue is not whether a [factfinder] rationally could have acquitted but whether it rationally could have found guilt beyond a reasonable doubt." United States v. Seng Tan, 674 F.3d 103, 107 (1st Cir. 2012). And read correctly – again, in a way most agreeable to the judge's guilty finding – the testimony illumined above about the multistate wire-communications process for direct deposit completely undoes Tum's very speculative suggestion that not a single transmission crossed state lines. He tries to play down this direct-deposit testimony, suggesting it only highlights how the direct-deposit process works generally, not how it worked here. But we think a clear-sighted factfinder could conclude otherwise.
Next, Tum insists that the government stumbled by not proving that the interstate nature of the wire transmissions was reasonably foreseeable to him. He is off base with this one too. The wire-fraud statute's interstate-nexus requirement is purely jurisdictional and not a substantive element of the offense, meaning the government did not have to prove that Tum had reason to think that a communication would cross state lines. Instead, all [3] that the government had to show was that it was reasonably foreseeable to a person in Tum's shoes that a wire communication would be made to further the scheme. See, e.g., Woodward, 149 F.3d at 63-64; Fermín Castillo, 829 F.2d at 1198-99. And there was plenty of evidence of that: For openers, a significant number of communications he got from Unum came either directly from its South Carolina benefits center or from its Maine-based office but with the benefits center's Palmetto State address emblazoned on them. Also, no one can deny that electronic communications go on all the time today, see United States v. Mullins, 613 F.3d 1273, 1281 (10th Cir. 2010), particularly in the world of banking, see, e.g., Fermín Castillo, 829 F.2d at 1198-99. So having asked that benefit [4] payments be deposited electronically into his account, a sensible person in Tum's position should have anticipated that wire transmissions involving Unum and others would follow – at least a reasonable factfinder using "common sense" could so conclude, which is all that is required. See Woodward, 149 F.3d at 64; see generally United States v. Ortiz, 966 F.2d 707, 712 (1st Cir. 1992) (holding that, in grading a sufficiency challenge, factfinders need not "divorce themselves from their common sense" or "abandon the dictates of mature experience"). Of course, the communication actually had to have crossed state lines, see generally United States v. Lewis, 554 F.3d 208, 213 (1st Cir. 2009) (explaining that the wire-fraud statute "require[s] actual crossing of a state . . . border"), but the government proved that in spades, as we pointed out above.
The unavoidable bottom line is that the evidence was strong enough to uphold Tum's wire-fraud convictions. We shift focus, then, to his complaint about the adequacy of the evidence underpinning his wire-fraud-conspiracy conviction – a complaint we likewise consider de novo, reading the record in the light most compatible with the government's theory of the case. See, e.g., Guerrier, 669 F.3d at 7. To put his claim in context, we start with some basic principles of conspiracy law.
A conspiracy is an agreement between the defendant and another or others with a particular kind of object – to commit a crime. See, e.g., United States v. Fenton, 367 F.3d 14, 19 (1st Cir. 2004). One must be a willing participant but need not "know[] the exact scope and extent of the collective endeavor." United States v. Ruiz, 905 F.2d 499, 506 (1st Cir. 1990). Knowing its essential nature suffices. See, e.g., United States v. O'Campo, 973 F.2d 1015, 1019 (1st Cir. 1992) (holding that "the government need not establish that the [members] knew or agreed upon every detail of the conspiracy," but only "the essential nature of the plan and their connections with it") (citation and internal quotation marks omitted)). Also, the government may prove an [5] agreement by circumstantial evidence – say, for example, by showing "a common purpose (such as a purpose to sell illicit drugs), overlap of participants, and interdependence of various elements in the overall plan." Martínez-Medina, 279 F.3d at 113-14. And a criminal factfinder need not turn a blind eye to what "is perfectly obvious." United States v. Ingraham, 832 F.2d 229, 240 (1st Cir. 1987).
As argued by the parties, everything here turns on the agreement element. A wealth of evidence shows that Tum and HHCS (the judge-found coconspirators) agreed to scam someone. Remember how right after Unum reminded Tum in December 2007 of his job/income-disclosure requirements (one of the umpteenth reminders, it seems) HHCS created a batch of false documents – contracts, invoices, checks, letters, tax-related papers, etc. – to hide his HHCS employment and pay. And remember too that Tum had a tax specialist use the false info in working up his e-filed joint tax return. But did the government have to prove that HHCS knew when it teamed up with Tum that the specific object of the conspiracy was a wire fraud against Unum?
Tum answers with an unequivocal "yes." And, he quickly adds, not a scrap of evidence shows that HHCS knew that he was trying to pull a fast one on Unum, which, he says, puts HHCS in the same boat as Hussen – whom the judge had acquitted. Even the government concedes that nothing directly shows that HHCS knew that Tum was defrauding Unum. To Tum's way of thinking, the evidence at most suggests that HHCS agreed to help him cheat the IRS, not swindle Unum. And because the duo did not agree on the Unum fraud, his theory continues, we must reverse the conspiracy conviction.
Hold on, the government says. All HHCS had to know was that Tum was perpetrating a fraud involving his income, not that Unum was the one being bilked. And given the volume of evidence on that score, the conspiracy conviction survives a sufficiency review.
Neither side cites any controlling caselaw that directly answers the question. But this is not quite the issue of first impression here that the parties make it out to be.
Helpfully, the Supreme Court has said in another context that the government in a conspiracy case need not prove the identity of a specific victim, unless the statute underlying the conspiracy charge so requires. See Feola, 420 U.S. at 672-73, 684, 686-93 (affirming a conviction for conspiring to assault a federal officer even though the defendant had no idea that the victim was a federal officer). So the key to the conundrum lies in the statutory language of the act that the judge found Tum and HHCS had conspired to infract – the wire-fraud act. As we said many pages ago, the core elements of wire fraud are a scheme to defraud involving an interstate communication by wire. Conspicuously missing from this list is any requirement that the schemers know the identity of the fraud victim. The same is true for the mail- fraud statute, see 18 U.S.C. § 1341, the elements of which mirror the wire-fraud statute in relevant respects, so cases dealing with one statute are helpful in dealing with the other. See, e.g., Carpenter v. United States, 484 U.S. 19, 25 n.6 (1987). Critically, one of our mail-fraud cases flatly rejected the idea that the government is obliged to prove that the defrauders intended to defraud a specific victim, reasoning that the mail- fraud act "requires only that there be a scheme to defraud . . . ." United States v. Royal, 100 F.3d 1019, 1030 (1st Cir. 1996) (citing, most relevantly, Sawyer, 85 F.3d at 723). We see no reason for a different result in wire-fraud cases. And because [6] the fraud victim's identity is not an essential element of wire fraud, neither is it an essential element of conspiracy to commit wire fraud. See, e.g., Feola, 420 U.S. at 686, 696 (holding that a conspiracy conviction requires proof of the same mental state as the substantive offense itself).
That conclusion torpedoes Tum's theory. Against this legal backdrop, the evidence spotlighted above, eyed in a prosecution-friendly way, was sufficient for a sensible factfinder to find that Tum and HHCS agreed on the essential nature of a wire- fraud conspiracy – i.e., that they agreed that Tum would defraud someone with an interest in his income by misrepresenting that Hussen worked for HHCS (keep in mind the checks from HHCS to Hussen, who never worked there, and also the false tax info that HHCS whipped up, which Tum based his e-filed joint-tax return on). Admittedly, the case is close. Yet we think that, applying the relevant law to the specific facts involved here, this agreement is enough to form the essential nature of the plan, even if HHCS did not know whether the misrepresentations would be used to defraud an insurance company, the IRS, or someone else.
In a last attempt to turn the tide, Tum contends, rather skillfully, that United States v. Rosenblatt, 554 F.2d 36 (2d Cir. 1977), should move us to reverse. We stand our ground.
Rosenblatt, a college dean, laundered through his college's bank account checks that his supposed co-conspirator, Morris Brooks, had handed him. Id. at 37-38. Brooks, a postal worker, had fraudulently induced the government to issue the checks. Id. Rosenblatt had no idea about Brooks's fraud – Brooks had said that the checks were legal but that he needed them laundered to help others evade taxes or conceal kickbacks. Id. at 38. Anyway, the government indicted the pair for conspiring to defraud the United States, in violation of 18 U.S.C. § 371, the general federal-conspiracy statute. See id. at 37. That statute punishes two things: conspiracy "to commit any offense against the United States," and conspiracy "to defraud the United States . . . ." Brooks pled guilty. Id. Rosenblatt rolled the dice and went to trial. Id. A jury found him guilty. Id. But the Second Circuit reversed, concluding that there was no evidence that the two had agreed to commit the same type of fraud on the United States – Brooks's fraud involved unlawfully obtaining checks, and Rosenblatt's involved tax evasion (for the most part). Id. at 37- 40.
Tum's case is different in at least two ways. For one thing, unlike Rosenblatt's, Tum's involves a conspiracy under the wire-fraud statute (see 18 U.S.C. §§ 1343 and 1349), not the rather "amorphous" conspiracy-to-defraud proviso of section 371 – a [7] proviso, by the way, that requires a showing of a specific target, i.e., the United States (or one of its agencies). For another [8] thing, as we have just explained, the evidence read in light of the guilty finding shows that Tum and HHCS agreed to do the same crime – wire fraud. Ultimately, then, Rosenblatt cannot save the day for Tum. And that is that.
Final Words Our work done, we affirm Tum's convictions in all respects.
NOTES
[1] Section 1343 punishes [w]hoever, having devised or intending to devise any scheme or artifice to defraud, or for obtaining money or property by means of false or fraudulent pretenses, representations, or promises, transmits or causes to be transmitted by means of wire . . . communication in interstate . . . commerce, any writings, signs, signals, pictures, or sounds for the purpose of executing such scheme or artiface . . . . And section 1349 punishes "[a]ny person who attempts or conspires to commit" wire fraud (among other offenses).
[2] Actually, to be fair, one of the argument headings in his opening brief accuses the government of not proving any of the key elements of wire fraud. But he spends all of his time there discussing the interstate-wire issue. He does say in his reply brief that he never "willfully" kept anything from Unum – rather, his difficulty with English prevented him from knowing what Unum expected of him, or so he claims. But because he floated this theory in his reply (not opening) brief, it is forfeited. See, e.g., Battista v. Clarke, 645 F.3d 449, 456 (1st Cir. 2011). And even if we were willing to overlook this problem, he would still lose, because on this record, viewed as it must be in the light most flattering to the verdict, see, e.g., Guerrier, 669 F.3d at 7, a levelheaded factfinder could conclude that Tum knew what Unum wanted him to do – something that leaps off the transcript pages, given his fessing up to his Portland Press Herald job but not to his more lucrative HHCS work after Unum employees explained what was required of him and how what he earned could affect his benefits.
[3] See, e.g., United States v. Lindemann, 85 F.3d 1232, 1241 (7th Cir. 1996); United States v. Blackmon, 839 F.2d 900, 907 (2d Cir. 1988); United States v. Bryant, 766 F.2d 370, 375 (8th Cir. 1985); cf. United States v. Robinson, 843 F.2d 1, 6 (1st Cir. 1988) (making a similar point in discussing a statute dealing with gun- toting felons, noting that the "passage of a gun in interstate commerce . . . can make a felon's receipt of that gun, later and without knowledge of or involvement in its previous interstate journey, illegal"); see generally United States v. Feola, 420 U.S. 671, 677 n.9 (1975) (stressing that "the significance of labeling a statutory requirement as 'jurisdictional' is . . . merely that the existence of the fact that confers federal jurisdiction need not be one in the mind of the actor at the time he perpetrates the act made criminal by the federal statute").
[4] See generally United States v. Muni, 668 F.2d 87, 90 (2d Cir. 1981) (stressing that "[t]he content of reasonable foreseeability must inevitably keep pace with advances in technology and general awareness of such advances").
[5] See also United States v. Sánchez-Berríos, 424 F.3d 65, 75 (1st Cir. 2005); United States v. Martínez-Medina, 279 F.3d 105, 113 (1st Cir. 2002); United States v. Martin, 228 F.3d 1, 11 (1st Cir. 2000); see generally United States v. Piper, 35 F.3d 611, 615 (1st Cir. 1994) (clarifying that to prove the required voluntary participation, the government must prove both "an intent to agree and an intent to effectuate the commission of the substantive offense").
[6] On this we are not alone. See, e.g., United States v. Howard, 619 F.3d 723, 727 (7th Cir. 2010); United States v. Munoz, 430 F.3d 1357, 1368-69 (11th Cir. 2005); United States v. Henningsen, 387 F.3d 585, 590 (7th Cir. 2004); United States v. Crawford, 239 F.3d 1086, 1093 (9th Cir. 2001); United States v. Loayza, 107 F.3d 257, 260-61 (4th Cir. 1997); United States v. Hatch, 926 F.2d 387, 392 (5th Cir. 1991).
[7] See United States v. Stavroulakis, 952 F.2d 686, 690-92 (2d Cir. 1992) (distinguishing Rosenblatt on a similar basis – i.e., that, unlike the Stavroulakis defendants, Rosenblatt was "charged under the inherently . . . amorphous 'conspiracy to defraud' clause of the general conspiracy statute" – and upholding convictions for money-laundering conspiracy when one conspirator thought the money came from drugs and the other thought it came from gambling).
[8] See United States v. Brandon, 17 F.3d 409, 421-22 (1st Cir. 1994) (discussing Tanner v. United States, 483 U.S. 107, 128-32 (1987)).