United States v. Vaknin

112 F.3d 579, 1997 U.S. App. LEXIS 10074, 1997 WL 217363

• Court: Court of Appeals for the First Circuit
• Decided: May 6, 1997
• Docket: 19-1960
• Status: Published

Opinion

SELYA, Circuit Judge.

These consolidated appeals raise, inter alia, an interesting question anent the standard of causation that courts must apply in fashioning restitutionary orders under the Victim and Witness Protection Act (VWPA), 18 U.S.C. §§ 3663(a), 3664(a) (1994). The appeals arise out of a multi-count indictment: each of the three appellants bribed the same bank official in connection with the making of loans; some of the loans soured; the bank failed; and the Federal Deposit and Insurance Corporation (FDIC) was left holding an empty bag. When the appellants pled guilty to criminal charges, the district court imposed sentences which included orders of restitution to cover what the court considered to be the attributable losses.

The appellants now challenge these impositions, and, in addition, one appellant, citing his cooperation with the prosecution, assails the district court’s refusal to depart downward from the guideline sentencing range (GSR). We affirm the court’s eschewal of a downward departure, uphold one restitutionary order (albeit with a modest modification), vacate the other two, and remand for further findings.

I. AN HISTORICAL PERSPECTIVE

Compulsory restitution as a societal response to criminal wrongdoing dates back over 4,000 years to the Code of Hammurabi and the Old Testament. See, e.g., Exodus 22:1-3 (“If a man shall steal ... he should make full restitution.”). In its earliest iterations, the practice was designed to forfend against the high social costs of blood feuds and the wreaking of personal vengeance by compensating victims in a more civilized way. See generally Thomas M. Kelly, Note, Where Offenders Pay for Their Crimes: Victim Restitution and Its Constitutionality, 59 Notre Dame L.Rev. 685, 686-88 (1984). By the Middle Ages, however, the sovereign had begun to administer the criminal law directly, and criminal restitution fell into desuetude. See id. The device remained moribund for several centuries. In the United States, for example, federal judges were not able to impose criminal restitution as a condition of probation until 1925 when Congress passed the Federal Probation Act, 18 U.S.C. § 3651 (repealed 1984). Even then, judges used the power sparingly. See Peggy M. Tobolowsky, Restitution in the Federal Criminal Justice System, 77 Judicature 90, 90-91 (1993).

The tectonic plates shifted in 1982 when Congress enacted the VWPA in response to a growing cognizance of victims’ rights. Notable for the speed of its election-year passage — the legislation was introduced in the Senate on April 22,1982, and signed into law by President Reagan less than six months later — the VWPA transmogrified criminal restitution from a sporadically imposed condition of probation into the sentencing norm in cases involving quantifiable economic loss.

The congressional purpose that animated the VWPA is no secret: “the court in devising just sanctions for adjudicated offenders, should insure that the wrongdoer make good[ ], to the degree possible, the harm he has caused his victim.” S.Rep. No. 532, at 31 (1982), reprinted in 1982 U.S.C.C.A.N. 2515, 2536. To accomplish this purpose, a district court, when pronouncing sentence, “may order, in addition to ... any other penalty authorized by law, that the defendant make restitution to any victim of such offense.” 18 U.S.C. § 3663(a). In determining whether to award restitution (and, if so, in what amount), the sentencing court “shall consider the amount of the loss sustained by any ■victim as a result of the offense, the financial resources of the defendant, the financial needs and earning ability of the defendant and the defendant’s dependents, and such other factors as the court deems appropriate.” Id. at § 3664(a).

In general, restitution under the VWPA is limited to “the loss caused by the specific conduct that is the basis of the offense of conviction.” Hughey v. United States, 495 U.S. 411, 413, 110 S.Ct. 1979, 1981, 109 L.Ed.2d 408 (1990). 1 When the fact, cause, or amount of the loss is disputed, the government must establish it by a preponderance of the evidence. See United States v. Baker, 25 F.3d 1452, 1454-55 (9th Cir.1994); United States v. Diamond, 969 F.2d 961, 967 (10th Cir.1992); see also 18 U.S.C. § 3664(d).

II. THE FACTUAL PREDICATE

We present the facts relevant to these appeals as best they have presented themselves, mindful that the record is noticeably underdeveloped.

Kenneth Annarummo was a bad apple. While working as a loan officer for Attleboro-Pawtueket Savings Bank (APSB or the Bank), he solicited and accepted bribes from numerous customers. Annarummo’s skulduggery came to light after the Bank failed and the FDIC intervened. In due course, the government indicted Annarummo and several eomplicit borrowers, including appellants Moshe Vaknin, Michael J. Fonseca, and E. Erie Yeghian (all real estate developers). 2 We recount the circumstances of each appellant’s involvement.

A. Vaknin’s Troubles.

Vaknin first approached APSB in 1987, seeking to refinance several properties. Informed by Annarummo that his request for funds would be facilitated if he greased the wheels, Vaknin paid Annarummo $17,500 and thereafter received the loan. In 1988, Vaknin sought to borrow more money and Annarummo again asked for a bribe in exchange for his assistance in getting the loan underwritten. Vaknin paid him $12,500 prior to securing loan approval. This sequence repeated itself later that same year, when Vaknin slipped Annarummo another bribe and secured a third loan (which was approved by the bank after a series of machinations in which Annarummo presented false information to the'credit committee). Although Vaknin repaid the initial refinancing in full, he defaulted on both the 1988 loans and the Bank sustained losses in excess of $900,000.

When indicted, Vaknin pled guilty to a single count of bank bribery. See 18 U.S.C. § 215 (1994). The Presentence Investigation Report (PSI Report) did not recommend restitution. In response to the prosecution’s-objection, the probation officer explained:

[I]t is not clear as to whether the losses incurred by the bank were a direct result of a fraudulent loan being negotiated as a result of the bank bribery or whether the losses were attributable to other factors, such as a downturn in the economy which affected the real estate market.

At the disposition hearing,' Judge Boyle sentenced Vaknin to an incarcerative term of twelve months and one day, two years’ supervised release, and a $50 special assessment. On the restitution issue, the judge sided with the prosecution; concluding that there would have been no funds advanced if the bribes had not been paid, the judge held Vaknin hable for the losses resulting from the defaulted loans, rejected the probation officer’s “downturn in the economy” hypothesis, and ordered Vaknin to pay restitution to the FDIC in the sum of $1,000,000.

B. Fonseca’s Troubles.

By the time Annarummo arrived on the scene, Fonseca was a valued customer of the Bank, having roughly $750,000 in outstanding loans. This debt had been incurred through normal channels and without subterfuge, mostly in connection with single-family residential properties in Rhode Island. Annarummo made no immediate demands on Fonseca, and Fonseca succeeded in securing additional financing through APSB.

In 1987, Fonseca encountered business difficulties and became fearful that he would not be able to meet the repayment schedule on an outstanding APSB note. When he voiced concern to Annarummo, the banker demanded a bribe for his help in warding off trouble should a default ensue. Fonseca paid Annarummo $3,000 but proved able to meet his payment obligation on time and in full.

In 1988, Fonseca applied for a $4,250,000 loan to cover the development of a much larger project than he had ever tackled — a subdivision of more than 50 lots in Bristol, Rhode Island. The record suggests (though it does not pin down) that, after approval of the loan request but prior to its disbursement, Annarummo demanded one of the lots as a bribe. Fonseca acquiesced and transferred title to Annarummo’s nominee, leaving one less lot as security for APSB’s loan.

The Bank terminated Annarummo’s employment in March 1990. Fonseca’s subdivision loan (which had a remaining principal balance of $611,500) was then 30 days in arrears, and Annarummo’s successor recommended foreclosure. Fonseca negotiated with APSB (which knew nothing of the bribes), and the parties agreed to enter into a forbearance agreement (FA) under which Fonseca would make a lumpsum payment of $450,000 in full satisfaction of the outstanding indebtedness. Fonseca tendered the funds within the agreed 35-day period. In time, the Bank failed, the FDIC intervened, the bribes were discovered, and the indictment materialized.

Fonseca pled guilty to a single count of bank bribery. The district court sentenced him to serve twelve months and one day in prison and a three-year term of supervised release. The court also imposed a $5,000 fine and a $50 special assessment. The matter of restitution proceeded much as in Vaknin’s case. The probation officer recommended against a restitutionary impost; the prosecution objected; and the district judge sustained the objection, ordering Fonseca to make restitution in the sum of $161,500 (the difference between the loan balance and the amount that Fonseca paid pursuant to the FA).

C. Yeghian’s Troubles.

Yeghian, a newcomer to APSB, applied for a loan of $2,930,000 in 1988 to fund the purchase of real property located in Providence, Rhode Island. Annarummo demanded a bribe of $20,000 (although the record is tenebrous as whether he approached Yeghian before or after the loan had been approved). In any event, Yeghian, using a corrupt lawyer as an internuncio, paid the bribe out of the loan proceeds.

Later that same year, Yeghian sought a loan of $1,400,000 to acquire and develop a parcel of real estate in Seekonk, Massachusetts. Once again, Annarummo demanded a bribe and received $22,909.52. 3 Both loans turned sour. The Bank’s demise, the FDIC’s entry onto the scene, the deterration of the bribes, and the indictment followed.

Yeghian pled guilty to one count of bank bribery. At sentencing, Judge Boyle imposed a ten-month prison sentence, a three-year supervised release term, a $10,000 fine, and a $50 special assessment. Rejecting a recommendation contained in the PSI Report, the judge ordered Yeghian to pay restitution in the sum of'$2,213,654.74.

III. THE DEPARTURE DECISION

Vaknin challenges the incarcerative portion of his sentence. The salient facts are as follows. The court sentenced Vaknin under the 1988 edition of the federal sentencing guidelines. The court figured the GSR as 8-14 months (adjusted offense level 11; criminal history category I), and this calculation is not in dispute. At the time of sentencing, the government asked the court to depart downward because Vaknin had made a good faith effort to render substantial assistance. See USSG § 5K1.1 (“Upon motion of the government stating that the defendant has made a good faith effort to provide substantial assistance in the investigation or prosecution of another person who has committed an offense, the court may depart from the guidelines.”). For his part, Vaknin solicited an even more generous departure. Nevertheless, departure decisions are entrusted primarily to the courts, and the sentencing judge’s role cannot be usurped by agreements between the prosecutor and the defendant. See United States v. Mariano, 983 F.2d 1150, 1154 n. 3, 1155-56 (1st Cir.1993). Exercising this authority, the court refused to impose a sentence below the GSR. Vaknin assigns error.

Vaknin’s claim of error is doubly flawed. The short, entirely dispositive answer to it is that he stakes out his position in a perfunctory manner. For that reason, the argument is deemed waived. See, e.g., United States v. Tardiff, 969 F.2d 1283, 1287 (1st Cir.1992); United States v. Zannino, 895 F.2d 1, 17 (1st Cir.1990).

The slightly longer but equally dis-positive answer is that, in the main, departure decisions are discretionary, and appellate review of refusals to depart is tightly circumscribed. See Koon v. United States, — U.S. -, —---, 116 S.Ct. 2035, 2046-47, 135 L.Ed.2d 392 (1996); Bruce M. Selya & Matthew Kipp, An Examination of Emerging Departure Jurisprudence Under the Federal Sentencing Guidelines, 67 Notre Dame L.Rev. 1, 13-14 (1991). Jurisdiction will only attach “when it appears that the failure to depart stemmed from the sentencing court’s mistaken impression that it lacked the legal authority to depart or, relatedly, from the court’s misapprehension of the rules governing departure.” Mariano, 983 F.2d at 1153. No such bevue occurred here.

To be sure, Vaknin labors to find a cognizable error. In this vein, he contends that the district court believed itself unable to depart downward because Vaknin had not provided information about his fellow borrowers’ criminal activities but only about the bribe-taker’s criminal activities. He builds this contention on scraps drawn from counsel’s colloquy with the judge at the disposition hearing. But an appellate court, seeking to ascertain a sense of what transpired at sentencing, must look to the whole of the record rather than isolated snippets extracted from it. See, e.g., United States v. Santiago, 83 F.3d 20, 25 (1st Cir.1996); United States v. Rostoff, 53 F.3d 398, 407 (1st Cir. 1995); cf. United States v. Tavano, 12 F.3d 301, 304 (1st Cir.1993). Applying this tenet here, the record, read as a seamless whole, belies Vaknin’s contention. ■

We need not tarry. The sentencing transcript shows with pristine clarity that Judge Boyle knew he could depart once the government invoked USSG § 5K1.1, but chose instead to impose a sentence within the GSR. As we read the record, his reasons for demurring were clear — and entirely permissible. In his view, Vaknin’s cooperation had been adequately rewarded because (a) the government had prosecuted only one count of bribery despite the fact that Vaknin had paid multiple bribes referable to separate borrowings, and (b) Vaknin’s offense level (and, hence, the GSR) already had been reduced for acceptance of responsibility under USSG § 3E1.1.

The transcript also reveals that the court weighed the quintet of factors under which a substantial assistance motion must be evaluated: the nature and extent of the assistance provided; its significance and utility to the prosecution; its timeliness; the truthfulness and reliability of the information conveyed; and the injury to, or risk exposure of, the defendant resulting from his cooperation. See Mariano, 983 F.2d at 1156 (enumerating factors and explaining that “[a], district court, faced with a section 5K1.1 motion, must at a bare minimum indicate its cognizance of these factors”). After mulling these and other relevant considerations, the court determined that, under the specific circumstances of Vaknin’s case, no departure was warranted. Such a decision is quintessentially a judgment call, and, thus, within the sentencing court’s discretion. See Tardiff, 969 F.2d at 1290. Consequently, we lack both the authority to second-guess the departure decision and the inclination to do so.

IV. THE CAUSATION QUANDARY

All three appellants challenge the district court’s restitutionary orders. Those challenges are similar insofar as they implicate the standard of causation. Therefore, we treat them in the ensemble to that extent.

A. Standard of Review.

Restitution orders customarily are reviewed under an abuse of discretion rubric. See United States v. Hensley, 91 F.3d 274, 277 (1st Cir.1996). In the course of this review, the sentencing court’s subsidiary factual findings must be credited unless they are clearly erroneous. See id. To the extent that a challenge to a restitution order hinges on a legal question, however, the sentencing court’s answer to that question is reviewed de novo. See United States v. Gilberg, 75 F.3d 15, 20 (1st Cir.1996); United States v. Savoie, 985 F.2d 612, 619 (1st Cir.1993). The appellants’ allegation that the district judge employed an improper legal standard of causation presents such a question.

B. The Legal Landscape.

The level of causation required under the VWPA is not immediately apparent, and the parties’ views on the subject are sharply divergent. The appellants advance a theory of “direct” causation, exhorting us to rule that restitution can be imposed only if the victim’s losses result directly from the offense of conviction and therefore that restitution cannot be imposed when an intervening phenomenon (e.g., a collapsing real estate market) is the more immediate cause of the loss. 4 Transposed into the métier of this case, the appellants’ theory seemingly would require the government to eliminate the possibility of concurrent causes and prove that the FDIC’s losses occurred as a direct result of the bribes that Annarummo solicited and received. The government cannot do so, the appellants posit, because stimuli unrelated to the bribes, such as intervening market forces, caused the ultimate losses.

The government’s counter-argument is that “but for” causation suffices; it urges us to rule that restitution can be imposed as long as the victim’s losses would not have eventuated but for the criminal activity. But for the bribes, this thesis runs, there would have been no loans, without which there would have been no losses. In this very general sense, the bribes caused the losses— and that, to the government’s way of thinking, is enough.

The appellants’ rejoinder is twofold. First, they debunk the legal standard articulated by the government. Second, they say that even if this articulation accurately reflects the state of the law, it does not justify the district court’s restitutionary orders. On the appellants’ shared hypothesis, the loans would have issued whether or not the bribes were forthcoming; thus, the Bank would have incurred the losses even if the appellants had played it straight.

The parties’ positions stand at opposite ends of a continuum. Our effort to determine where on the continuum the correct legal standard is housed starts with the language of the VWPA itself. Section 3663(a) authorizes restitution to “any victim” for a covered offense. This provision must be read in tandem with section 3664(a), which directs the sentencing court to consider “the amount of the loss sustained by any victim as a result of the offense.” For purposes of this case, see supra note 1, restitution is appropriate only for “the loss caused by the specific conduct that is the basis of the offense of conviction.” Hughey, 495 U.S. at 413, 110 S.Ct. at 1981.

Since the text of the VWPA does not speak explicitly to the dimensions of the requisite standard of causation, 5 we must consult other sources in our quest to discover it. Next on the list is legislative history. This material, like the statute itself, does not specifically limn the standard of causation. Nonetheless, it offers some important insights.

In enacting the VWPA, Congress strove to encourage greater use of a restitutionary remedy. See S.Rep. No. 532, supra, 1982 U.S.C.C.A.N. at 2536-37. At the same time, it disclaimed any intent to convert the main event — the sentencing hearing — into a time-consuming sideshow — prolonged litigation over restitution-related issues. This disclaimer was made manifest in a variety of ways. For example, rather than requiring great precision in fixing the amount of restitution due, Congress visualized the VWPA as “authoriz[ing] the court to reach an expeditious, reasonable determination of appropriate restitution by resolving uncertainties with a view towards achieving fairness to the victim.” Id. at 2537.

In short, the legislative history clearly signals a congressional preference for rough remedial justice, emphasizing victims’ rights. In our view, this preference counsels against importing a stringent standard of causation (such as might be appropriate in a tort context) into the VWPA.

Of course, rough remedial justice does not mean leaving matters to the whim of the sentencing judge, and Congress did not conceive of restitution as being an entirely standardless proposition. The government must bear the burden of establishing the loss, 18 U.S.C. § 3664(d), and an award cannot be woven solely from the gossamer strands of speculation and surmise. See United States v. Neal 36 F.3d 1190, 1200-01 (1st Cir.1994). By like token, just as insisting upon a modicum of reliable evidence reinforces the specific advantages of the restitutionary remedy, so too does insisting upon a certain degree of causal precision. As the Supreme Court has noted, demanding a “direct relation between the harm and the punishment gives restitution a more precise deterrent effect than a traditional fine.” Kelly v. Robinson, 479 U.S. 36, 49 n. 10, 107 S.Ct. 353, 361 n. 10, 93 L.Ed.2d 216 (1986).

Finding the legislative history suggestive rather than compelling, we examine the case-law. In previous decisions, this court has remarked the broad policy goals of the VWPA and concluded that difficulty in achieving an exact measurement of victim loss should not preclude the imposition of restitution. See Savoie, 985 F.2d at 617. On the subject of causation, however, our decisions have tended to involve either situations in which the closeness of the causal link could not seriously be questioned, see, e.g., United States v. Lilly, 80 F.3d 24, 28 (1st Cir.1996), or those in which we found restitution to have been ordered in contravention of Hughey, see, e.g., United States v. Newman, 49 F.3d 1, 11 (1st Cir.1995). Neither polar extreme brings much light to the vexing issue which these appeals present.

Neal is the only .notable exception to this taxonomy. That case featured a defendant who had been found guilty both of being an accessory after the fact to a bank robbery and of laundering funds. The district court imposed a restitutionary award that equalled the bank’s entire loss from the thievery. We vacated the award, noting that it could not be determined on the sparse record available “whether the court calculated, pursuant to Hughey, the portion of [the bank’s] losses that were actually caused by the specific criminal conduct forming the basis for Neal’s convictions.” 36 F.3d at 1200 (italics omitted). We instructed the district court, on remand, to hold a hearing on the causation issue and modify the award to the extent that any portion of the loss was not “attributable to” NeaPs criminal conduct. Id. at 1201. In dictum, we cautiously suggested that some varietal of but for causation might suffice. See id. at 1201 n. 10 (“If ... evidence is presented indicating that Neal played a significant role in helping the other defendants escape and that but for his actions, there was a substantial likelihood that the full proceeds would have been recovered, the court could well be within its statutory authority in imposing the full [restitutionary amount].”). Thus, circuit precedent furnishes a weak indication that but for causation can suffice under the VWPA.

Reading the out-of-circuit cases is like attending a bar association meeting in a small town; one can find congenial cases, like friendly faces in the crowd, to support almost any standard of causation for the VWPA. We have found decisions which appear at least superficially to reject but for causation in favor of a “direct result” standard. See, e.g., United States v. Silkowski, 32 F.3d 682, 689-90 (2d Cir.1994); Ratliff v. United States, 999 F.2d 1023, 1026-27 (6th Cir.1993). By contrast, we have found decisions which seem to accept unqualified but for causation as sufficient under the VWPA. See, e.g., United States v. Keith, 754 F.2d 1388, 1393 (9th Cir.1985); United States v. Richard, 738 F.2d 1120, 1122-23 (10th Cir.1984). We have found decisions which straddle the question, see Government of the Virgin Islands v. Davis, 43 F.3d 41, 46 (3d Cir.1994) (seemingly endorsing, in a single paragraph, both but for and direct causation), and those which confess confusion on the issue, see United States v. Cloud, 872 F.2d 846, 856 n. 13 (9th Cir.1989) (acknowledging “a conflict in this circuit regarding the nexus the government must establish between the defendant’s criminal conduct and the victim’s losses to support a VWPA restitution order”).

C. Choosing a Standard.

Upon close perscrutation, the extreme positions advocated by the parties do not hold out much promise in our quest for a serviceable standard of causation.

On the one hand, the sort of direct causation standard that the appellants propose is simply too rigid. Under their theory of intervening forces, a court could not impose restitution even if the defendant’s conduct were a substantial cause of a loss, unless it were the last cause. Such a standard would flout the basic purpose of the VWPA. 6 In our judgment, Congress did not contemplate such adamantine formalism when it moved to expand the availability of restitutionary remedies by enacting the VWPA. See S.Rep. No. 532, supra, 1982 U.S.C.C.A.N. at 2537.

On the other hand, concerns of fairness require us to reject the unbridled but for causation standard that the government propounds. Under it,' a court could impose restitution based on the most tenuous of connections. 7 While it is true that for want of a nail the kingdom reputedly was lost, cf. Benjamin Franklin, Poor Richard’s Almanac (1758), it could hardly have been Congress’ intent to place the entire burden on the blacksmith if the nail was an insignificant factor in the calculus of concurrent causes. Such a result would countervail principles of fundamental fairness and, in the bargain, would be at odds with the majority of reported cases. See, e.g., United States v. Holley, 23 F.3d 902, 914-15 (5th Cir.1994); United States v. Tyler, 767 F.2d 1350, 1351-53 (9th Cir.1985).

Having rejected the parties’ proposals, it falls to us to fashion the appropriate legal standard. Despite the gaps in the statute and in its legislative history, and notwithstanding the contradictions that permeate the cases, we think it is possible to distill certain bedrock principles from the sources that we have consulted.

” First: Restitution should not be ordered in respect to a loss which would have occurred regardless of the defendant’s conduct. A good illustration of this principle in operation is found in United States v. Blackburn, 9 F.3d 353 (5th Cir.1993). There, the sentencing court included foreclosure expenses in calculating the amount of restitution due. The Fifth Circuit reversed, citing proof that the foreclosure would have happened even if the defendant had not committed the crime. See id. at 359; see also United States v. Walker, 896 F.2d 295, 305-06 (8th Cir.1990) (holding that when defendants, who owned a company, defrauded the United States, restitution to laid-off company employees was improper because the record failed to show that the fraud caused the company to cease operations).

Second: Even if but for causation is acceptable in theory, limitless but for causation is not. Restitution should not lie if the conduct underlying the offense of conviction is too far removed, either factually or temporally, from the loss. We offer two examples of remoteness in fact. The first arises in a ease that bears some similarity to the instant case.

In Diamond, 969 F.2d at 963-64, the defendant pled guilty to filing false financial reports with a lender. The loan had already been made before Diamond authored the reports, but the reports apparently helped in obtaining an extension. The loan proved uncollectible. The sentencing court ordered the defendant to make restitution, reasoning that the loss stemmed from the false reports. The court of appeals refused to equate the extension of an existing loan with the granting of the loan in the first place, and negated the restitutionary order because there was no proof that the extension worsened the lender’s position. See id. at 966.

A somewhat different example of factual remoteness is found in United States v. Sablan, 92 F.3d 865 (9th Cir.1996). There, the defendant had been convicted of computer fraud. The district court ordered restitution for expenses incurred by the victim in meeting with investigators to discuss the case. The Ninth Circuit struck these amounts from the award, ruling that the expenses were not connected closely enough to the fraudulent conduct. See id. at 870; see also United States v. Kenney, 789 F.2d 783, 784 (9th Cir.1986) (invalidating that portion of a restitution- order which was designed to reimburse the corporate victim for the cost of having its employees testify at the defendant’s trial, but upholding that part of the order encompassing the cost of removing film chronicling the robbery from the bank’s surveillance cameras).

Typical of the situations in which but for causation existed but restitution was denied because the claimed losses were temporally remote is Holley, in which the court deemed restitution improper when the victim, who received foreclosure property from the defendant in the course of the criminal activity, unnecessarily held onto the property for a lengthy interval after the crime was discovered, and the property declined in value during that period. 23 F.3d at 914-15. Similarly, in Tyler, the defendant cut down a tree in a national forest and was apprehended as he tried to take it to a nearby lumber mill. 767 F.2d at 1351. The government retained the lumber, needlessly, for a long period of time, then sold it in a fallen market for considerably less than it would have fetched if sold promptly. See id. The district court ordered restitution, pegging the loss by reference to the reduced price. The appellate court disagreed, pointing out that, although abstract but for causation existed, it was too attenuated to support the award. See id. at 1351-53.

Consistent with these two principles and with our dictum in Neal, 36 F.3d at 1201 & n. 10, we hold that a modified but for standard of causation is appropriate for restitution under the VWPA. This means, in effect, that the government must show not only that a particular loss would not have occurred but for the conduct underlying the offense of conviction, but also that the causal nexus between the conduct and the loss is not too attenuated (either factually or temporally). The watchword is reasonableness. A sentencing court should undertake an individualized inquiry; what constitutes sufficient causation can only be determined case by case, in a fact-specific probe.

D. Applying the Standard.

Having elucidated the appropriate legal standard, we turn finally to the causation questions embedded in the appeals that are before us. These appeals, like the decisions canvassed above, provide some insights into the standard’s operation.

1. In Vaknin’s case, restitution is appropriate. The district court specifically found that the bribes which Vaknin paid were a but for cause of the Bank’s losses on the defaulted loans. The record contains no basis on which to mount a credible challenge to this finding. After all, the arrangements for the bribes preceded the making of the loans, and the bribes were admittedly paid in exchange for Annarummo’s assistance in procuring the loans.

Moreover, common sense must inform inquiries into restitution under the VWPA. See S.Rep. No. 532, supra, 1982 U.S.C.C.A.N. at 2536-37. In Vaknin’s ease, the evidence clearly shows not only that the loans were procured by bribery but also that the bribe-taker connived to bend the rules; in at least one instance Annarummo shaded the presentation to APSB’s credit committee to increase the likelihood that the loan would be forthcoming. We believe that where, as here, the government establishes that arrangements for a bribe precede and relate to the making of a loan, a commonsense inference arises that subsequent losses referable to the loan’s uneolleetibility are causally linked in reasonable proximity to the bribe. Cf, e.g., Blinzler v. Marriott Int'l Inc., 81 F.3d 1148, 1158-59 (1st Cir.1996) (discussing commonsense inference that arises from proof that a relevant document has been destroyed); United States v. Olbres, 61 F.3d 967, 971-72 (1st Cir.) (discussing commonsense inference that arises in tax evasion case from proof of expenditures in excess of declared income and disposable assets), cert. denied, — U.S. -, 116 S.Ct. 522, 133 L.Ed.2d 430 (1995). Of course, the inference can be rebutted if the defendant produces specific evidence of factual or temporal remoteness. Here, however, Vaknin made no such showing. To the contrary, there is no compelling proof either of an unforeseeable intervening cause or of any cognizable remoteness, factual or temporal.

That ends the matter. Because the record adequately supports Judge Boyle’s finding of but for causation, and contains no sufficient suggestion of factual or temporal remoteness, restitution for the losses resulting from the tainted loans is altogether appropriate.

2. We treat Fonseca’s and Yeghian’s appeals in tandem. In both instances, the record is so exiguous that the very existence of but for causation seems problematic. As to Fonseca, the single loan in respect to which the court ordered restitution may have been approved by the Bank independent of, and prior in time to, Annarummo’s solicitation of a bribe. 8 On the present record, we simply cannot tell — and the lower court made no specific finding on the point. The question is potentially important because, if it turns out that the Bank approved the loan prior to any arrangements for a bribe, then in such event the circumstances would not support an inference of but for causation; and, in the absence of such an inference, it is difficult (although not impossible) to conceive how a sufficient causal link between bribe and loss could be forged.

Moreover, the record suggests that even if but for causation exists, the requisite connectedness might be lacking. Fonseca argues with some force that the Bank’s loss, if one occurred at all, was occasioned by its need for an immediate cash infusion; that this exigency gave birth to the FA; and therefore, no cognizable loss occurred. 9 But this argument, too, depends on facts which the record does not contain, and on which the lower court made no particularized findings. It is clear that Fonseca’s loan was overdue and that the Bank had a right to call the loan. From that point forward, it is unclear whether the Bank entered into the FA merely as a quick fix for its own problems or because it wanted to mitigate an inevitable loss.

As to Yeghian, the record is similarly inexplicit about the timing of his arrangements with Annarummo vis-á-vis the Bank’s approval of the subject loans. There is some indication that one (if not both) of the loans on which he defaulted may have been approved independent of' any deal with Annarummo, but the sentencing court made no detailed findings and the extant record is too sparse to permit us to answer the causation questions with confidence.

It would be unprofitable to delve more deeply into these matters. We are confronted by a largely undeveloped record, embellished with few 'specific findings. Given that enigmatic reality, remand is required. We envision that the district court, the next time around, will direct the parties to augment the record with respect to (a) the presence or absence of a causal link between Fonseca’s and Yeghian’s criminal conduct and the FDIC’s losses, (b) if that causal link is demonstrated, the closeness of the connection, factually and temporally, between that conduct and the ultimate losses, and (c) such other matters as the court may deem suitable. We anticipate further that the court will make particularized findings on each disputed issue. We intimate no view as to the proper outcome of those proceedings.

V. MISCELLANEOUS

Three final matters require brief attention. The first is a matter raised by Fonseca and Yeghian. The others relate solely to Vaknin’s obligations.

A. Picking up the Tab.

It is apodictie that restitution only can be ordered to redress a loss to a victim. See United States v. Gibbens, 25 F.3d 28, 33 (1st Cir.1994). Using this truism as a lever, Fonseca and Yeghian question whether the VWPA allows the court to order restitution to the FDIC for losses originally sustained by the (now failed) Bank. The question is easily answered.

Following existing circuit precedent, we hold that the benefit of the VWPA’s remedial provisions extends to a government agency which, in the exercise of duly delegated powers, steps into the shoes of the original victim. See id. at 32-35. Thus, if a failed bank was a victim of the defendant’s criminal activity, the FDIC, as its insurer and receiver, itself qualifies as a victim for purposes of a restitutionary order under the VWPA. See United States v. Phaneuf, 91 F.3d 255, 265 (1st Cir.1996).

B. Ability to Pay.

Vaknin argues that the district eourt abused its discretion by ordering him to make restitution' without considering his ability to pay. We agree with Vaknin’s premise that judicial consideration of a defendant’s ability to pay is statutorily mandated as a prerequisite to an order for restitution. See 18 U.S.C. § 3664(a). We disagree, however, with his conclusion that the lower eourt neglected to touch this base.

We have stated with a regularity bordering on the monotonous that the consideration requirement does not mean that a judge must decide the question in a particular way or even that he must make express findings on the record as to the defendant’s ability to pay. See, e.g., Newman, 49 F.3d at 10; Savoie, 985 F.2d at 618. It is enough if “the record on appeal reveals that the judge made implicit findings or otherwise ade quately evinced his consideration” of this factor. Savoie, 985 F.2d at 618.

Here, the PSI Report spelled out Vaknin’s past earnings history and current financial condition in appreciable detail. The sentencing transcript indicates that the judge absorbed this information, voiced his skepticism about Vaknin’s ability to comply with the restitution order as matters stood, 10 but nonetheless impliedly found that a sufficient possibility of eventual repayment existed. We think that this finding is supportable. A defendant’s impoverishment today is no assurance of future poverty, and, hence, present impecuniousness is not a bar to the imposition of restitution. See United States v. Brandon, 17 F.3d 409, 461 (1st Cir.1994); United States v. Lombardi, 5 F.3d 568, 573 (1st Cir.1993). A sentencing court permissibly may take into account a defendant’s earning capacity and the prospect that his fortunes will improve. See Lombardi 5 F.3d at 573; Savoie, 985 F.2d at 619.

Here, the judge apparently issued a restitution order as a hedge against his founded belief that the defendant — an individual of demonstrated entrepreneurial bent — might well acquire assets in the future. While this conclusion would have been less controversial had the judge made a more pointed reference to Vaknin’s past accomplishments and future financial prognosis, we cannot say that an abuse of discretion transpired. See Lombardi 5 F.3d at 572-73.

C. The Government’s Concessions.

The district court ordered Vaknin to make restitution in the amount of $1,000,000. This figure is vulnerable on two fronts. First, the government has brought to light on its own initiative a mathematical error that, when corrected, will reduce the amount of restitution owed. 11 Second, the sentencing court premised the loss calculation on the amount which the Bank received when it resold the property Vaknin had pledged to secure the defaulted loans, rather than on its fair market value at the time of foreclosure. Because the district court used fair market value as of the foreclosure date when determining the amount of restitution that Yeghian owed, the government concedes that it would be fair to employ the same barometer in respect to Vaknin (a similarly situated codefendant). We accept the government’s concessions at face value, without passing substantively upon them, and direct the district court to make these two adjustments to the restitutionary award. The resultant obligation thus will be reduced to $902,000.

VI. CONCLUSION

We need go no further. For the reasons set forth herein, we affirm the convictions of all the defendants; modify the restitution order imposed against Vaknin, and, as modified, affirm it; vacate the restitution orders imposed on Fonseca and Yeghian, respectively; and remand for further proceedings as to them.

Affirmed in part; vacated in part; remanded.

1 . The defendant in Hughey had used credit cards in an unauthorized manner, and the Court limited restitution to the loss attributable to the lone count on which he had pled guilty (as opposed to the total loss from all his fraudulent conduct). Congress reacted by amending the VWPA in November of 1990, adding § 3663(a)(2) [the former § 3663(a) became § 3663(a)(1), but its substance remained essentially unchanged]. This amendment provides that “a victim of an offense that involves as an element a scheme, a conspiracy, or a pattern of criminal activity means any person directly harmed by the defendant’s criminal conduct in the course of the scheme, conspiracy, or pattern.” As we explained in United States v. Hensley, 91 F.3d 274, 276-77 (1st Cir.1996), restitution for all criminal conduct done in the course of a single scheme, conspiracy, or pattern of activity is now appropriate, whether or not the defendant has been convicted of (or even charged with) the specific acts, as long as the offense of conviction has as an element the broader scheme, conspiracy, or pattern.

There are two reasons why the 1990 amendment has no bearing here. In the first place, the criminal conduct of which the appellants stand convicted occurred prior to the date of the amendment. Thus, the pre-1990 version of the VWPA governs our inquiry. See United States v. Royal, 100 F.3d 1019, 1032 (1st Cir.1996); United States v. Gilberg, 75 F.3d 15, 20-21 (1st Cir. 1996). In the second place, the offenses of conviction here do not have as an element any broader scheme, conspiracy, or pattern.

2 . Annarummo eventually pled guilty to three counts of bank bribery, 18 U.S.C. § 215 (1994), and one count of subscribing to a false tax return, 26 U.S.C. § 7206(1) (1994).

3 . The odd amount stems from the fact that the bribe took the form of a payment by Yeghian to liquidate an outstanding loan encumbering Annarummo's Porsche.

4 . While the appellants profess to know direct causation when they see it, they have been unable either to muster a comprehensive definition of the term or to suggest a viable limiting principle. The government's arguments in support of but for causation, see infra, suffer from much the same vice.

5 . Though the amended version of the VWPA does not apply to this case, see supra note 1, the appellants asseverate that the amendment’s use of the adverb "directly" heralds Congress' intent vis-á-vis the type of causation that it envisioned. We do not agree. The legislative history of the 1990 amendment plainly indicates that the language employed, albeit containing the word "directly,” does not support the appellants’ theory of causation. As Congress explained:

The use of "directly” precludes, for example, an argument that a person has been harmed by a financial institution offense that results in a payment from the insurance fund because, as a taxpayer, a part of a person’s taxes go to the insurance fund.

H.R.Rep. No. 681(1), at 177 n. 8 (1990), reprinted in 1990 U.S.C.C.A.N. 6472, 6583 n. 8. This definition ranges far afield from the definition of direct that the appellants tout. Thus, we conclude that the 1990 amendment did not alter the standard of causation applicable to VWPA cases.

6 . Imagine a situation in which D, a convicted felon who is carrying a handgun, is speeding down a highway, fleeing from the authorities. D's car slams into an unregistered automobile, with defective brakes, owned and operated by Stranger (S), causing S to swerve and hit V, who suffers severe injuries. D is then prosecuted for reckless endangerment and found guilty. S’s miscreancy should not preclude a court from ordering D to make restitution for V’s medical expenses. Yet the appellants’ theory would erect just such a barrier.

7 . Imagine a situation similar to that described in note 6, supra; but, instead of being prosecuted for a vehicular offense, D is charged with and convicted of being a felon in possession of a handgun. While but for causation may be present — after all, but for his unlawful possession of a weapon, D would have had no occasion to flee from the authorities, and, thus, would not have been careening down the road and would not have precipitated the accident — it is hard to make a principled argument that the offense of conviction (felon in possession) supports an order against D to make restitution for V’s medical expenses. Yet the government’s theory ordains just such a result.

8 . Fonseca’s past lending relationship with the Bank tends to support this inference. It suggests, at the least, that the Bank considered him creditworthy well before Annarummo hove into view.

9 . In substance, Fonseca asserts that by accepting an accelerated $450,000 payment under a consensual pact (the FA) in satisfaction of the outstanding loan balance ($611,500), APSB did no more than make a business judgment designed not to salvage a failing loan — Fonseca says he could have paid it off in full, given time — but to shore up a failing bank. On that basis, he reasons that APSB (and ultimately the FDIC) suffered no loss.

10 . Indeed, the judge explicitly declined to levy a fine against Vaknin, noting on the judgment form that no fine would be imposed due to an inability to pay.

11 . This is very professional behavior, and we commend the prosecutors for it.