SENTENCING OPINION

GERARD E. LYNCH, United States District Judge

On October 2, 2002, Hakeem Mohammed pled guilty to mail fraud and conspiracy to commit mail fraud, in connection with a scheme to steal credit information and use it to obtain credit cards, convenience checks, and other lines of credit in the names of unsuspecting victims. Determining the appropriate sentence involves a number of issues under the Sentencing Guidelines.

FACTS

In June 2002, FBI agents investigating the theft of credit reports of thousands of customers of the Ford Motor Company learned that GMAC Mortgage Corporation had recently received a change-of-address application from one of its credit customers, changing the address on the customer's line of credit from an address in Troy, Michigan, to an address in the Bronx. Inquiry of the customer disclosed that he had neither moved nor sent the change-of-address form, and further investigation revealed that he was one of the Ford customers whose credit information had been stolen. A mail cover on the Bronx address soon identified mail addressed to another Ford victim not known to reside at that address. The mail was traced to a bank, which advised the agents that the a line of credit had recently been opened with the bank in the name of the addressee; needless to say, in actual fact that person had neither moved to the Bronx nor requested credit from the bank. Other similar mail directed to the Bronx address involving fraudulently obtained credit cards or lines of credit was soon identified.

In early July, the agents made a controlled delivery of convenience checks issued in the name of one of the victims to the Bronx address. The person residing there quickly cooperated with the authorities, and told them that the defendant Hakeem Mohammed paid her to receive packages addressed to various individuals, and that she had been doing so for several months. On July 29, 2002, when Mohammed came to pick up the mail, he was surveilled, and eventually arrested, by FBI agents. When arrested, Mohammed had in his possession not only the mail from the Bronx address, but also a list of names of other individuals with addresses, telephone numbers, credit card account numbers and expiration dates, and dollar amounts. Mohammed claimed that he had picked up mail from the Bronx address at the request of a man he knew as "Saheed," who has never been further identified. In addition to this arrest, Mohammed had earlier been arrested by local authorities in Westchester in connection with the same scheme. At the time of that arrest, three different women gave written statements to the police indicating that they had received merchandise using fraudulent credit information that derived from Mohammed; one of them also admitted to permitting her address to be used by Mohammed in connection with the receipt of credit documents pursuant to the scheme.

SENTENCING ISSUES

This case exemplifies some of the difficulties involved in attempting to quantify the seriousness of varying types of fraud by a set of pre-determined rules. To assess the seriousness of crimes of fraud and theft, the guidelines rely primarily on the amount of loss occasioned by the crime. The underlying intuition is sensible, and works reasonably well for simple thefts: If we imagine two bank officers, for example, each with unlimited access to their employers' funds, one who embezzles $100,000 clearly deserves more punishment than one who embezzles only $10,000. But across a wide range of diverse types of schemes and thefts, the amount of the loss may be harder to compute, and, once computed, may be a less accurate proxy for the seriousness of the offense. The guidelines include a number of complex provisions, some at issue in this case, in an effort to capture at least some of the factors that bear on this assessment. In light of the multifarious character of schemes and frauds, the effort to develop a precise set of rules that will capture all variations of culpability is no doubt futile, and courts must be alert to apply the rules in a way that reflects the underlying purposes of sentencing, and to identify variations in culpability not adequately specified in the guidelines that might require a departure from the mathematical calculation of the guideline sentence. Loss Calculation. Although the base offense level for mail fraud is only 6, it is subject to rapid escalation based on the amount of loss. U.S.S.G. § 2B1.1. In a scheme such as Mohammed's, however, the correct calculation of loss that can be determined from the evidence known to the authorities is neither intuitively obvious nor obviously indicative of the seriousness of the offense.

Mohammed argues that the amount of loss attributable to him should be found to be at most $16,986, which is the total of the charges actually made to credit devices in the names of victims identified either from mail directed to addresses which he controlled or from documents on his person when he was arrested. The government, on the other hand, argues that the correct calculation is $267,945, which is the total of the available credit limits on the various accounts set up in the names of those victims.

It should be apparent that neither of these amounts represents a precise calculation of the harm caused by Mohammed and his co-conspirators, or of the seriousness of the offense. First, it is a virtual certainty that the score or so of identity theft victims whose credit information is known to the FBI to have come into Mohammed's possession represent the tip of a much larger iceberg. Neither the credit charges actually incurred nor the larger total of available credit represents anything like a precise calculation of the extent of the scheme in which this defendant was involved. Second, neither number makes any pretense of measuring the actual harm inflicted by this crime, including the damage actually or potentially inflicted on the reputations and credit ratings of the affected individual victims; the time, effort and expense incurred by those individuals and the credit agencies involved to unravel the false information created by the conspirators; or the damage done to the trust essential to commercial relationships in an economy in which credit plays such a large part. No simple calculation of dollar "loss" will adequately measure the seriousness of this crime. Third, both amounts are in an important way arbitrary, and reflect the policies of lenders and the effectiveness of law enforcement authorities as much as or more than the magnitude of the crime. Within the limited universe of this set of identified victims, the amounts actually charged to various accounts were presumably only a beginning; there is no reason to believe the conspirators intended to stop extracting money from these accounts until the credit limits were reached. But even the full extent of the available credit lines understates the ambitions of the schemers. Presumably, with access to the credit histories of many victims, Mohammed and his co-conspirators were able to select victims with credit records that would support high credit limits. The credit limits established, however, reflect the policies of various credit issuers as much as or more than they reflect the goals of the criminals. Moreover, Mohammed had in his possession the names and credit information for a number of individuals for whom accounts had not yet been opened. There is every reason to believe that the conspirators intended to open accounts or obtain credit using this information. The credit limits on the accounts that had already been opened thus considerably understate the losses he and his co-conspirators intended to inflict.

With these limitations in mind, then, we turn to the rules set forth in the guidelines for the calculation of loss. Mohammed points to Application Note 2 Fail to U.S.S.G. § 2B1.1, the portion of the commentary that deals most directly with calculating loss in cases of credit card fraud. That Note provides:

In a case involving any counterfeit access device or unauthorized access device, loss includes any unauthorized charges made with the counterfeit access device or unauthorized access device and shall be not less than $500 per access device.

Accordingly, Mohammed contends that the amount of loss should be the actual amounts charged to the credit devices, or $16,986.

Mohammed's argument is implicitly premised on the assertion that Note 2(F)(i) sets forth a controlling rule that limits the loss calculation in credit card fraud cases. The language of the Note, however, does not clearly support this view. The Note provides that the loss "includes" unauthorized charges, not that the loss is limited to such charges. That is, rather than defining the appropriate loss calculation, it merely requires the inclusion of certain items in the calculation, thus setting a floor under the loss calculation in credit fraud cases. The only limiting language is that which sets a bottom limit to the loss, providing that the loss "shall be not less than $500 per card."

The language of others of the "Special Rules" set forth in Note 2(F) confirms that the Sentencing Commission knew how to define a loss calculation when it chose to do so. Thus, in Note 2(F)(vi), the Commission provides that when illegal drugs are stolen, the "loss is the estimated street value" of the drugs, and in Note 2(F)(vii), regarding stolen cultural artifacts, the "loss . . . shall be determined in accordance with" a particular formula. (Emphasis added.) Other rules, like that providing the $500 floor in Note 2(F)(i), use mandatory language with respect to aspects of the calculation. Although the introduction to the list of "Special Rules" indicates that they are to be used "[n]otwithstanding" the general principles of Application Note 2(A), it is clear that only certain of the rules (like those for drugs and cultural artifacts) displace those principles by providing an alternative definition of loss. Others presuppose the application of those general rules, except to the extent that the "Special Rule" specifically supersedes or supplements them. For example, the rule for Ponzi schemes. Note 2(F)(iv), makes clear that the loss "shall not be reduced by" the gains made by some investors, but provides no rule for calculating the loss itself, evidently presupposing that the general rules of Note 2(A) will govern the situation. Thus, there is no reason to assume that the "Special Rules" are intended in all cases to displace the general principles of loss calculation. As the introduction to those rules states, they are to be "used to assist in determining loss in the cases indicated." Note 2(F) (emphasis added). The particular relationship to the general rules depends on the language of each particular special rule.

Note 2(F)(i) in particular does not appear to be intended to displace the general rule that "loss is the greater of actual loss or intended loss." Application Note 2(A). For example, if a defendant attempted to purchase a $6000 item with a stolen credit card, but was apprehended after an alert clerk refused to accept the card, the "intended loss" of $6000 would surely be the appropriate measure of loss; the loss would not be limited to $500 because no "unauthorized charges" were actually "made with the unauthorized access device." The limitation for which Mohammed argues is similarly inappropriate for the fraud at issue here. As the government points out, a criminal with a stolen credit card may have no idea how much he will be able to charge before the card is cancelled, and a counterfeit card by definition may have no set limit at all. In such instances, the credit limit theory urged by the government in this case would be inapplicable, and the charges actually "made," required to be "include[d]" in the loss by Note 2(F)(i), would most likely be the outer limit of the loss. In this instance, however, Mohammed apparently had knowledge of the credit limits on each account, since he or his co-conspirators had opened each of them; in some instances he had in his possession notes listing the credit limits on individuals' accounts. The logical inference is that he and his co-schemers anticipated milking the accounts up to the available limits. On this basis, the government and the Probation Office correctly argue that the Court should find a loss of $267,945.

Whatever might be the case in other instances of credit card fraud, under the particular circumstances revealed in this case, the Court finds that the defendant and his co-conspirators intended to inflict losses as near as possible to the credit limits on the accounts they opened or the lines of credit they obtained. Other courts have reached similar results. A District Judge with particular expertise in sentencing, for example, has held that "where a sentencing court has facts upon which to base findings that a defendant was capable of and intended to use [access devices] to secure amounts at or near their credit limits, aggregating the credit limits of the cards to calculate loss is appropriate." United States v. Say, 923 F. Supp. 611, 614 (Vt. 1995) (Sessions, J.). Accord, United States v. Egemonye, 62 F.3d 425 (1st Cir. 1995); United States v. Sowels, 998 F.2d 249, 250 (5th Cir. 1993). Accordingly, the Court finds that the amount of loss in this case is conservatively estimated at $267,945. Under U.S.S.G. § 2B1.1(b)(1)(G), this requires an upward adjustment of twelve levels, to level 18.

Credit Cards. There is no dispute that the offense level must be further adjusted upward by two levels because the offense involved trafficking in access devises. U.S.S.G. § 2B1.1(b)(9)(B). The offense level, to this point, is thus 20.

Number of Victims. In its letter pursuant to United States v. Pimentel, 932 F.2d 1029, 1034 (2d Cir. 1991), the government expressed its belief that the offense level should be further enhanced because the offense involved more than ten victims, pursuant to U.S.S.G. § 2B1.1(b)(2)(A)(i). The Probation Office rejected this adjustment, however, citing Application Note 3(A)(ii)(I), which defines a "victim" for purposes of this guideline, as "any person who sustained any part of the actual loss determined under subsection (b)(1)." According to the Probation Office, the "actual loss" in this case — presumably the actual charges made on the various access devices in question — was suffered not by the 20 or so individuals whose stolen credit information was in Mohammed's possession, or even by those whose credit was actually utilized, but by the seven merchants or financial institutions that ultimately bore the losses from these charges. (Presentence Investigation Report at 13-14.) Both the defendant and the government now agree that with this analysis.

The Application Note is not a model of draftsmanship. The Note's reference to "the actual loss determined under subsection (b)(1)," for example, is at best infelicitous. Section 2B1.1(b)(1) does not require a determination of actual loss. Rather, it refers to "loss," which, Application Note 2(A) tells us, is the greater of "actual loss" or "intended loss" as defined in Application Notes 2(A)(i) and 2(A)(ii), respectively. Nor is it obvious why the Sentencing Commission meant (if it did) to restrict the enhancement for multiple victims to cases where actual loss was suffered.

The "multiple victim" enhancements were adopted as part of the revision and consolidation of the theft and fraud guidelines in 2001. They replaced or incorporated two prior guideline sections, formerly part of the fraud guideline. One, which is retained in § 2B.1(b)(2)(A)(ii), provided a two-level enhancement for frauds "committed through mass-marketing." See former U.S.S.G. § 2F1.1(b)(3) (2000). The other provided a two-level enhancement for offenses that involved "more than minimal planning" or "a scheme to defraud more than one victim."See former U.S.S.G. § 2F1.1(b)(2) (2000). Neither the mass-marketing nor the "more than one victim" enhancement were limited to cases in which multiple victims experienced actual loss. By definition, the mass-marketing enhancement dealt with the means that were used to solicit the many targeted victims, not with the number of targets, if any, as to whom the scheme was successful. See U.S.S.G. § 2B 1.1, Application Note 3(A)(i), formerly U.S.S.G. § 2F1.1, Application Note 3 (2000) (defining "mass-marketing"). A "scheme to defraud more than one victim" was defined as "a design or plan to obtain something of value from more than one person," and a "victim" as "the person or entity from which the funds are to come directly." Former U.S.S.G. § 2F 1.1, Application Note 4 (emphasis added). Thus, both of the former enhancements were focused, logically enough, on the scope of the scheme, not on the number of victims who suffered actual loss.

The 2001 amendments retained that focus as to "mass-marketing," but apparently changed it as to the multiple victim enhancement. No explanation was provided. The Commission simply noted that the amendments deleted the "more than one victim" enhancement along with the related "more than minimal planning" concept, and provided instead graded enhancements for offenses involving ten or more victims, or 50 or more victims. "This provision is designed to provide a measured increment . . . for offenses involving larger numbers of victims," reported the Commission, noting that it applied to "those cases in which victims, both individuals and organizations, sustain an actual loss under subsection (b)(1)." Guidelines Manual, App. C, Amendment 617, at 1209 (2002). While this comment provides a persuasive reason to replace the "more than one victim" enhancement with graded enhancements for larger numbers of victims, it gives no reason for counting as victims only those who suffered actual financial loss, and excluding schemes that unsuccessfully targeted multiple victims or imposed non-economic harm on numerous persons.

This case provides a good example of a situation in which the number of victims who suffered actual loss inadequately measures the scope of the crime. The defendant had access to credit information or access devices affecting a score of individuals; it is a reasonable inference that he and his co-schemers intended to use that information to obtain property, at the expense of those individuals or the financial institutions to whom credit applications had been or would be made, or who had provided the individuals with credit that the schemers could access. Like a mass-marketed scheme, the effort here was to defraud multiple victims, even if quick action by the FBI prevented the schemers from inflicting "actual loss" on more than seven financial institutions.

The guidelines as written do not "adequately take into consideration" the aggravating factor that the scheme here, like a mass-marketing scheme, was aimed at multiple victims, even if multiple victims had not yet been successfully bilked, and thus an upward departure is potentially appropriate. 18 U.S.C. § 3553(b). Whether to exercise the Court's discretion to depart on that ground, however, should be considered after the guideline calculation is complete, and not during the calculation of the guideline sentence; to do otherwise risks placing "the departure cart before the Guidelines Range horse." United States v. Milan, 304 F.3d 273, 296 (3d Cir. 2002), quoted with approval in United States v. Jeffers, 329 F.3d 94, 101 n. 6 (2d Cir. 2003). For now, it is sufficient to note that the current guidelines do not provide for a mandatory upward adjustment on this ground.

Role in the Offense. The government and the Probation Office argue that an adjustment of four levels is appropriate because Mohammed was "an organizer or leader of a criminal activity that involved five or more participants." U.S.S.G. § 3B1.1(a). Mohammed disputes both his role in the offense and the extent of the organization.

Here too, the shadowy nature of criminal activity, and the limited information available to the authorities (and thus to the Court) bedevil the effort to reduce sentencing to mathematical precision. The government has submitted credible evidence that Mohammed obtained credit information for a couple of dozen of the thousands of victims whose credit information was stolen from Ford, and that at least four others were involved with him in the offense. He recruited and managed at least two of those individuals. Thus, at a minimum, the criminal activity involved five known participants, and Mohammed was a manager or supervisor in the operation. Any reasonable assessment of the scope of the scheme and the role Mohammed played in it would yield the conclusion that the scheme was wide-ranging, and likely involved a number of participants, and that Mohammed played a significant part in it. The guidelines, however, by their nature, require more specific factfindings and invite hair-splitting about whether Mohammed falls on one side or the other of finely-drawn lines.

First, the Court must determine the number of participants in the scheme. It is very likely that the number of people known to the authorities in connection with Mohammed's activities is just the tip of the iceberg. No one, including Mohammed, contends that he stole the Ford credit information or knows who did, and it is likely that the information has since spread widely and been used in many places by many people. Nevertheless, sentencing cannot be based on such speculation, and we must turn to the specific evidence presented to the Court.

The identities of four participants are undisputed: Mohammed himself, the elusive Saheed, and the two women, one in the Bronx and the other in Westchester, who confessed to the authorities that Mohammed paid them to permit the use of their addresses in connection with the fraud. To find a fifth, the government points to two other women, arrested in Westchester, who also gave statements to the police. As Mohammed points out, neither of these women even claims to have met him.

Nevertheless, even crediting Mohammed's claim that he did not know or recruit these women, the Court finds that at least one of them, Tahisha McGaha, is properly counted as a participant in the criminal activity in question. McGaha told the police that a friend of hers (the Westchester woman whose address was admittedly used by Mohammed) told her about the scheme, and offered to let her order merchandise from Neiman Marcus using a credit card in a name that the authorities independently linked to Mohammed. The merchandise was in fact ordered, and the package went to the home of the third woman, where McGaha signed for it, using a false name. The government argues that it can be inferred from the statement that Mohammed authorized this purchase by McGaha in an effort to recruit her to receive credit cards or merchandise. Although the statement is susceptible to that inference, it is not necessary to go so far. Even assuming that it was McGaha's friend, not Mohammed, who recruited her, and that Mohammed never used or attempted to use McGaha's services, McGaha's statement is clear that her friend told her that she had told Mohammed about McGaha and that he had authorized her to use one of his credit cards to order the goods.

By ordering merchandise for herself using the stolen credit information, McGaha clearly became a member of the conspiracy to misuse the credit information. Whether or not McGaha ever met Mohammed, and whether or not she ever provided other services to the conspirators, McGaha engaged in criminal conduct by using the fraudulent credit account. It is irrelevant whether this was permitted as an effort to recruit McGaha, or as a favor to her friend. What matters is that McGaha became part of the agreement to use the stolen credit information. She is thus a participant in the criminal activity at issue here.

Determining whether Mohammed is a manager/supervisor or an organizer/leader is a subtler question. The guidelines instruct that in addressing this question, courts should consider "the exercise of decision making authority, the nature of participation in the commission of the offense, the recruitment of accomplices, the claimed right to a larger share of the fruits of the crime, the degree of participation in planning or organizing the offense, the nature and scope of the illegal activity, and the degree of control and authority exercised over others." U.S.S.G. § 3B 1.1, Application Note 4.

The information available to the Court is inadequate to assess most of these factors with great confidence. As noted above, it is likely that more individuals were involved than those identified by the government. Information about thousands of people was stolen, and there is no evidence that Mohammed personally accomplished that theft. It is unknown whether others obtained that information and utilized it as Mohammed did, and if so whether they were employees who worked for some central leader or independent contractors who separately purchased information about limited numbers of victims from the original thief. Mohammed claims to have been working for one "Saheed," but nothing is known about this figure. We are largely in the dark about whether there was a leadership level (or levels) above Mohammed, whether he was the organizer or leader of a specific, limited local operation, or whether he played a broader leadership role in the theft and distribution of credit information to other criminals. That Mohammed recruited and controlled at least those he paid to receive his mail is clear; whether he played some part in conceiving or controlling the larger scheme, and the extent of his decision-making authority or share of the proceeds, is not.

The distinction between an "organizer/leader" and a "manager/supervisor" is clearly a matter of degree. The fact that Mohammed is the highest-ranking member of the scheme to have been caught does not necessarily mean that he is an organizer or leader of the crime. Those that he supervised are at the lowest levels of the organization; the existence of one or even more levels of authority higher than his in the hierarchy is entirely consistent with the evidence regarding his role in the offense. While the Sentencing Commission has made clear that in distinguishing these levels of authority, "titles such as `kingpin' or `boss' are not controlling," U.S.S.G. § 3B 1., Application Note 4, certainly it remains important to preserve a distinction between those who may exercise even greater levels of authority and creativity and those who play intermediate managerial roles. Given the limitations on the available evidence, the Court finds that it has been established by a preponderance of the evidence only that Mohammed was a "manager or supervisor" of criminal activity involving five or more persons. His offense level therefore must be increased by 3, to level 23.

Obstruction of Justice. In proceedings following his arrest, in an effort to obtain bail release. Mohammed gave his name as "Hakeem Mohammed" and gave a social security number. The Probation Office, however, developed evidence that he has told others that his real name is "Mutiu Lateef," and that name appears on his son's birth certificate. (Presentence Investigation Report at 11.) Moreover, the government has demonstrated that the social security number in question belongs to a different "Hakeem Mohammed," nearly 20 years older than the defendant, and that the defendant Mohammed is admittedly an illegal alien not entitled to any social security account. The government has thus proved by a preponderance of the evidence that Mohammed lied to the Magistrate Judge in connection with bail proceedings by providing at a minimum a false social security number. The government argues, logically enough, that this false statement was deliberately made in an effort to obtain bail release by covering up his status as an illegal alien. See United States v. Kirsch, 54 F.3d 1062, 1073-74 (2d Cir. 1995).

Further investigation of the facts, however, reveals that Mohammed apparently admitted his status as an illegal alien to the pretrial services officer from the very first interview. The government retreats to the dubious claim that Mohammed admitted only that he had entered the country illegally in 1997, not that he was still undocumented at the time of the arrest, and that the social security number was intended to convey the impression that he was now a legal resident. The Court rejects that argument. First, the record is insufficient to establish the actual words used by Mohammed; the only evidence available is the pretrial service officer's report to the Court, which is at best a paraphrase of whatever Mohammed said. Second, even if the report is taken as authoritatively recording Mohammed's words, there is no reason to believe that Mohammed intended to engage in the hair-splitting the government attributes to him. Third, there is no evidence that Mohammed ever claimed to any judicial officer that he was a legal resident of the United States.

It is tempting to conclude that giving a false name and social security number to a court determining bail is intrinsically an obstruction of justice, because such deception interferes with the court's ability to identify the defendant and make judgments about his ties to the community and the danger he may present. Our Court of Appeals has repeatedly emphasized, however, that false statements standing alone are insufficient to warrant this enhancement; a specific intent to obstruct justice is required. United States v. Brown, 321 F.3d 347, 351-52 (2d Cir. 2003); United States v. Bradbury, 189 F.3d 200, 204-05 (2d Cir. 1999). In this instance, the evidence that "Hakeem Mohammed" is not Mohammed's true name is weak; the government has not proved by a preponderance of the evidence that the name is false. While the social security number is unquestionably not his, there is no real reason to think that this was presented to the Court to throw the Magistrate Judge off the scent; it is more likely than not that the name and identification documents are simply those the defendant has regularly used in the United States. No specific intent to obstruct justice has been shown.

Acceptance of Responsibility. The government does not dispute that defendant is entitled to a three-level reduction for timely acceptance of responsibility pursuant to U.S.S.G. § 3E1.1. His total offense level is therefore 20.

Criminal History Category. The defendant's criminal history score is I. The government contends that "the criminal history category does not adequately reflect the seriousness of the defendant's past criminal conduct or the likelihood that the defendant will commit other crimes." U.S.S.G. § 4A1.3. In support of this claim, the government cites the fact that the defendant was "pending trial . . . on another charge at the time of the instant offense," a fact recognized by the guidelines as potentially supporting a departure on this ground, id. § 4A1.3(d), and that the scheme in question had been going on for months before his arrest. (Letter of AUSA Harry A. Chernoff to the Court, at 8-9 (May 16, 2003).) On the facts of this case, however, no departure is warranted. The prior state arrest on which the government relies was, as the government agrees, in connection with the very same scheme for which he is to be sentenced here. Thus, all of the facts relied on by the government go not to the severity of the defendant's prior criminal conduct, for none has been shown, but to the scope of the instant criminal scheme. The criminal history category, and departures based on its inadequacy as a predictor, deal with an assessment of the defendant's prior history, not with the seriousness or scope of the present criminal acts. However serious the present scheme is, its gravity does not alter the fact that the defendant has no prior record of other criminal activity.

Accordingly, with an offense level of 20 and a criminal history category of I, the guideline sentencing range is 33-41 months.

Downward Departures. In addition to the proposed bases for departures already considered and rejected, Mohammed argues that the Court should depart downward based on his willingness to waive removal proceedings and consent to deportation. He appears to argue that his immigration status justifies a departure (1) because such a waiver would save the government the burden and expense of deportation proceedings, (2) because his deportation will increase the severity of his punishment by separating him from his infant citizen child, and (3) because as an illegal alien, he would be subject to harsher conditions of confinement than a citizen who had committed the same offense. (Madden Letter at 5-6.)

A defendant's alienage is a factor that can be considered in assessing the appropriateness of a departure. United States v. Restrepo, 999 F.2d 640, 643-44 (2d Cir. 1993). However, it is "not ordinarily relevant" to sentencing, and in order to justify a departure a court must identify factors occasioned by the defendant's immigration status that are specifically relevant to sentencing and that have not been taken into account by the sentencing guidelines. Id. at 644.

Neither of Mohammed's first two arguments will support a departure. A departure based on a defendant's willingness to consent to deportation is permissible only where the defendant has a colorable, nonfrivolous defense to deportation, such that his consent to removal assists the administration of justice.United States v. Galvez-Falconi, 174 F.3d 255, 260 (2d Cir. 1999). Without such a defense, the departure is not permitted.United States v. Sentamu, 212 F.3d 127, 136-37 (2d Cir. 2000). Mohammed does not contend that he has any colorable argument against deportation, or present any reason to believe that his waiver of deportation proceedings would benefit the government or significantly expedite his removal from the United States. Nor do the hardships attendant on deportation warrant a departure. United States v. Adubefour, 999 F.2d 639 (2d Cir. 1993); Restrepo, 999 F.2d at 646-47.

The Second Circuit has also rejected departure based on the unavailability of preferred prison assignments to deportable aliens. Restrepo, 999 F.2d at 644-46. Several other courts of appeals, however, have upheld departures where a defendant's immigration status would subject him to harsh conditions of confinement, see United States v. Farouil, 124 F.3d 838 (7th Cir. 1997); United States v. Charry Cubillos, 91 F.3d 1342 (9th Cir. 1996); United States v. Smith, 27 F.3d 649 (D.C. Cir. 1994). Both the Seventh and Ninth Circuits argued that the Supreme Court's decision in United States v. Koon, 518 U.S. 81 (1996), undermined the logic of Restrepo. Farouil, 124 F.3d at 846-47; Charry Cubillos, 91 F.3d at 1344.

Whether Koon undermines Restrepo is for the Second Circuit to decide; this Court remains bound by its authority. See Tsang v. United States, No. 97 Civ. 1886 (CSH), 1997 WL 630182 (S.D.N.Y. Oct. 9, 1997) (recognizing binding authority ofRestrepo despite circuit conflict). But even if Restrepo did not bar departures on this ground, this Court would decline to depart in this case. First, even those courts that have permitted departure for unduly harsh conditions of confinement due to alienage have emphasized that such a departure would only be warranted where the conditions in question are "substantially more onerous than the framers of the guidelines contemplated in fixing the punishment range for the defendant's offense [ . . .,] and the differences in the conditions of confinement or other incidents of punishment between deportable aliens and other citizen (or nondeportable alien) defendants . . . are not great." United States v. Guzman, 236 F.3d 830, 834 (7th Cir. 2001). Accordingly, those courts agree that legitimate departures on these grounds will be "quite rare." Smith, 27 F.3d at 655. Ineligibility for half-way houses or minimum security institutions, the only consequences Mohammed relies upon, are not such extraordinary deprivations as to warrant a finding that the Commission did not take into account the chance that someone in this sentencing range would be subjected to them.

Second, and ultimately more importantly, after taking into account all of the circumstances of the case, the Court finds that the sentencing range set by the guidelines is not unduly harsh, even taking into account the additional burdens imposed on defendant due to his alienage, in light of the severity of the offense conduct. Accordingly, the Court would decline to depart downward in any event.

Upward Departures. Two grounds for possible upward departure remain to be considered. First, citing United States v. Karro, 257 F.3d 112 (2d Cir. 2001), the government argues that an upward departure of two levels is appropriate because the non-economic harms inflicted on the individual victims of the "identity-theft" scheme is not adequately taken into account by the economic measures of harm that dominate the guideline calculation. (Chernoff Letter at 4-6.) As the Karro court held, "the use of the names and social security numbers of multiple individuals . . . to open and obtain money from credit card accounts at banks and a department store clearly gave the District Court a sufficient basis to find a risk of substantial harm to those individuals" and thus to warrant a departure.Karro, 257 F.3d at 122.

Second, as discussed above, the Sentencing Commission's limited upward adjustment for multiple victims has not adequately taken into account the possibility, present in this case, that an offense that does not cause actual economic loss to ten or more victims may nevertheless have harmed more than ten people in non-economic ways, and been designed to inflict actual economic loss on more victims (fortuitously prevented by effective law enforcement intervention), such that the moral desert and need for deterrence that justify the adjustment apply equally to such an offense.

Either of these factors, in the Court's view, could warrant an upward departure of two levels (the same order of magnitude as the multiple-victim adjustment), if the sentence determined by the guidelines were inadequate to accomplish the sentencing goals dictated by Congress in the Sentencing Reform Act. Overall, however, on the facts of this case, the Court finds that the cumulative weight of the various adjustments already required results in a sentencing range that is adequate to permit the selection of an appropriate sentence. The severity of these adjustments, along with the undoubted additional sufferings resulting from the defendant's alien status, might well have resulted in a sentence at the low end of the sentencing range, absent the aggravating factors under discussion. Those factors can thus be given sufficient weight in the selection of an appropriate sentence within the range. The Court notes, however, that if any of the upward adjustments found above were not applicable, thus reducing the guideline offense level, the Court would depart upward on the basis of the number of victims and the extent of the non-economic harm to those victims. The sentence imposed by the Court is the sentence the Court finds appropriate, taking into account all the facts and circumstances of the case.

Plea Bargaining Considerations. Finally, Mohammed argues that the government's sentencing positions should be rejected because it would be inappropriate to arrive at a sentencing level that was near or above the statutory maximum, which would effectively deny him credit for his guilty plea and acceptance of responsibility. (Madden Letter at 2.) This argument, however, cannot affect the Court's assessment of the appropriate sentence. It may be in the government's interest to moderate its sentencing advocacy in order to give a defendant an incentive to plead guilty, and if a defendant will be sentenced at the statutory maximum in any event, the incentive provided by the acceptance of responsibility provisions of the guidelines may be defeated. But that is a concern for the government in plea negotiations. The responsibility of a sentencing court is to determine the appropriate punishment to impose in light of the guidelines and those considerations that legitimately permit departures, not to second-guess or compensate for exercises of prosecutorial discretion. United States v. Bonnet-Grullon, 212 F.3d 692 (2d Cir. 2000). Mohammed has received appropriate credit for acceptance of responsibility in the course of the Court's guideline calculation. If other appropriate sentencing factors nevertheless result in a sentence near the statutory maximum, the fact that the guilty plea discount is less than the defendant may have expected is not a basis for imposing a lesser sentence.

SENTENCE

Taking into account all the factors appropriate in assessing the seriousness of the defendant's conduct, the various aggravating and mitigating circumstances, and the purposes of sentencing, the Court concludes that a sentence of 41 months' imprisonment is appropriate in this case. SO ORDERED.