concurring.
Of the myriad challenges facing the Sentencing Commission when it drafted the Sentencing Guidelines, few were greater than the task of fashioning clear and sensible rules for the grouping of offenses. While what emerged in U.S.S.G. § 3D1.2 is a generally sensible and workable set of principles, this case suggests to me that facets of the grouping Guidelines are forced, and result in formalistic rather than common sense dispositions that yield results contrary to the Commission’s intent to limit the significance of the charging decision. While I join in Judge Sloviter’s carefully written opinion for the court, and in the judgment, I write separately to suggest that the grouping guidelines would benefit from a redraft that would elevate substance and common sense over form. My comments thus pertain only to Part III of the opinion of the court. I note that, from its earliest days, the Commission has urged the federal judiciary to make suggestions for Guideline revision, viewing them as a means of implementing the ongoing monitoring process. See U.S.S.G. Ch.l Pt. A *182§ 4(b) (stating that Commission will analyze judicial decisions to determine how to refine Guidelines); see also United States v. Woods, 24 F.3d 514, 518 n. 4 (3d Cir.1994) (discussing same). To its great credit, the Commission has frequently acted upon such suggestions.
I
I see the case as follows. Prior to the offenses of conviction, Rudolph was known by the government’s informant, Felix, to have sold stolen government property in return for bribes.' The events leading to both the bribery offense and the sale of government property (i.e. theft) offense at issue involved an offer to pay a sum of money in return for goods stolen from the Justice Department. In the instance leading to the theft conviction, Felix made the illicit offer. In the instance leading to the bribery conviction, Felix facilitated an offer made by a third-party, Wesley Clement. In both cases: (1) Rudolph and his buyer haggled over the amount of the bribe; (2) Rudolph used his position with the INS to facilitate the transaction; and (3) the end result was precisely the same—a theft of government property in return for a bribe.
Under the standard for § 3D1.2(d) grouping announced in United States v. Ketcham, 80 F.3d 789, 796 (3d Cir.1996), grouping is appropriate if: (1) the defendant’s offense behavior has been found to be continuous or ongoing, and (2) the offense guidelines take his ongoing course of conduct into account. See § 3D1.2(d) (offense guidelines must be “written to cover” ongoing offense behavior). Both the district court and this court have assumed that Rudolph’s-behavior was ongoing and continuous, see Op. at n. 3, and the facts supporting the charged offenses, .combined with the fact that Rudolph admitted to two other (uncharged) bribes in return for INS templates on two other occasions, suggest that this assumption is correct.
As previously noted, grouping pursuant to subsection (d) also requires that the offense guidelines “already take into account the fact that there has been a course of harmful conduct.” Ketcham, 80 F.3d at 796. The court concludes that neither § 2B1.1 nor § 2C1.1, the relevant offense guidelines here, explicitly takes a continuing course of conduct into account, and hence that Rudolph’s offenses should not be grouped. While the court has correctly applied the law, I find this outcome extremely troubling. ’ The express purpose of the grouping guidelines is to limit the significance of the formal charging decision and to prevent multiple punishment for substantially identical offense conduct, while still ensuring incremental punishment for significant additional conduct. See U.S.S.G. Ch. 3 Pt. D, intro, comment. The problem with § 3D1.2(d) is that, in spite of these goals, it requires that the offense guidelines be written to cover ongoing offense behavior. In other words, if we read the Guidelines literally, we could, as intimated in Ketcham, supra, be forced to reject a grouping because of how the offense guidelines are written when the reality of the offense conduct clearly shows substantially identical conduct. This seems to be an unnecessary and untoward exercise in formalism, with the added vice of effectively bringing the charging process back to the center of the sentencing calculus.1
That is what happened here. In my view, the key criminal behavior underlying both of Rudolph’s offenses is the theft of government property in return for bribes. The only difference is the particular item Rudolph was asked to steal from the Justice Department closet. Yet, because one offense was charged as a bribe and sentenced under § 2C1.1, and the other offense was charged as a sale of government property and sentenced under § 2B1.1, the offenses are held not to be groupable.2 Thus, it is the charging *183decision and not the actual offense conduct that is ultimately controlling the sentence here.
The court correctly states that we have not been given the authority to group offenses “carte blanche ” in order to “override a prosecutor’s decision to charge certain offenses and not others,” and that we are limited to the textual dictates of § 3D1.2. Op. at 9-10, 11. The court’s statements, however, point out precisely the tension that currently exists between the policy behind the grouping notion, the purpose of which is, I reiterate, to limit the significance of the formal charging decision and to prevent multiple punishment for substantially identical offense conduct while still ensuring incremental punishment for significant additional conduct, and the actual directions that we are given in the Guidelines themselves. Not only are we not given power to override the charging decision, but under § 3D1.2(d) as presently drafted, we are powerless to look beyond the charging decision because we must adhere to the text of the offense guidelines—and not the actual offense conduct—to decide if grouping is appropriate. What the foregoing analysis suggests is that the grouping guidelines suffer from a drafting flaw, for this result is not, I submit, what the Commission really intended. Neither is it good policy if that was the intended result.
II
This functional type of approach is also necessary to avoid analogous pitfalls arising under our current construction of § 3D1.2(b). Under that subsection, multiple counts are groupable when they “involve the same victim” and constitute “part of a common scheme or plan.” See § 3D1.2(b). According to the guideline application notes, when there is no identifiable victim, the “victim” for purposes of a § 3D1.2(b) analysis is “the societal interest that is harmed.” See id,., appl. note 2. In the present case, the court has rejected Rudolph’s subsection (b) claim (i.e. that his theft and bribery offenses were part of a common scheme and involved the same victim) based on the conclusion that the primary societal interest affected by “Rudolph’s acceptance of the bribes was the integrity and efficacy of the nation’s immigration policies,” whereas the victim of the PSI theft offense was the individual with respect to whom the PSI was prepared. See Op. at 181. As we are reviewing here for plain error, see Op. at 181, I cannot fault this result. The court’s subsection (b) victim analysis, however, suffers from the same reliance on formalisms that hampers the subsection (d) analysis.
There are at least two ways to look at a bribery offense when we try to assess who is the “victim” of the crime. On the one hand, we could examine the nature of the individual who accepted the bribe. On the other hand, we could also examine the nature of the task or property that was exchanged as consideration for the bribe. In this case, the court has chosen to assess solely the nature of the consideration to determine who or what is the “victim.” Not surprisingly, the court concludes that the stolen INS template affects the integrity of the immigration laws while the PSI affects the individual who is its. subject. ■ But if we had looked instead to the nature of the bribe acceptor, our analysis would have been quite different. In both Rudolph’s bribery and sale of government property offenses, we would find a Justice Department official who jeopardized the integrity of his agency by accepting bribes. Viewed in that light, in both instances the harm caused would be the undermining of the citizenry’s trust in government and its officials. Indeed, the district court noted that this was a common ground between Rudolph’s offenses. See App. at 69-70.
The unfortunate result of the court’s analysis is that a defendant like Rudolph who is charged with stealing one template and one PSI (even though there was evidence that Rudolph had stolen another template) is not eligible to have his offenses grouped, whereas if he happened to be charged with stealing the two templates, his offenses could have *184been grouped under subsection (b). By limiting our analysis to one side of the bribery equation, our interpretation of the Guidelines enforces formalism over the plain reality that there was one criminal scheme and it just so happened that the defendant was caught stealing a template and a PSI rather than two templates or two PSIs.
Moreover, application note 2 to § 3D1.2 also states:
[T]he “victim” for purposes of subsections (a) and (b) [in non-identifiable victim cases] is the societal interest that is harmed. In such cases, the counts are grouped together when the societal interests that are harmed are closely related---- Ambiguities should be resolved in accordance with the purpose of this section as stated in the lead paragraph, i.e., to identify and group “counts involving substantially the same . harm.”
This application note demonstrates the Commission’s recognition that the determination of what interest is harmed will necessarily be analytically imprecise. I believe, therefore, that this note directs us to rely ultimately not on a metaphysical parsing of what elements of society were harmed, but, so long as the societal interests are closely related, on whether the real elements of the offense suggest that there is only one-ultimate harm embodied in both counts. See United States v. Riviere, 924 F.2d 1289,1305 (3d Cir.1991) (applying earlier version of application note 2 and basing grouping decision on § 3D1.2 policy grounds). While I agree that the subject of the PSI is properly identified as a victim, I believe that because there has been a theft of Justice Department property, the government and the attendant societal interests are also implicated as victims by the sale of the PSI, and that a less formalistic approach could therefore suggest that grouping is appropriate here.
My approach is supported by a recent Seventh Circuit decision addressing a related “same victim” question. In United States v. Wilson, 98 F.3d 281 (7th Cir.1996), the defendant pled guilty to charges of money laundering and mail fraud arising out of a Ponzi scheme. Relying on a line of authority that proposed that the victim of mail fraud is the person defrauded, while the victim of money laundering is society at large, the government argued that these two offenses were inappropriate for grouping under subsection (b) because they involved different victims, and thus ■ different harms. See Wilson, 98 F.3d at 283. Although finding the government’s contention to be correct in the abstract, the court disagreed that this theoretical victim analysis compelled the conclusion that grouping was inappropriate:
[W]hen the defendant is convicted of laundering the proceeds of his fraud ... as Wilson was here, there is intuitive force to the argument that the victim of the fraud is also a victim of the transaction designed . to hide or “cleanse” the funds of which she was defrauded.
Id. (internal, citation omitted). I believe the Seventh Circuit’s decision to apply its “intuitive force” rationale in these circumstances once again demonstrates that we must look to the actual offense conduct and the nature of the criminal scheme involved—and not just rely on abstract victim analysis or other formalisms—to give meaning to the Commission’s intent behind the grouping guidelines.
Ill
The animating purpose of the grouping rules is to identify and recognize when multiple counts of conviction “are so closely intertwined with other offenses that conviction for them ordinarily would not warrant increasing the guideline range,” and to sanction accordingly. See U.S.S.G. Ch. 3, Pt. D. intro, comment; see also Wilson, 98 F.3d at 282 (basic goal of § 3D1.2 is to combine offenses involving closely related counts). To impose sentences consistent with this purpose, federal sentencing judges must be able to assess the factual realities of the conduct underlying the convictions to determine when the counts are related enough to demand grouping. Section 3D1.2(d), as currently written, and § 3D1.2(b), as currently construed, stand in the way of this process, and the sentence imposed on Deandre Rudolph bears this out.
Therefore, in addition to the revised construction of subsection (b) proposed supra, I also suggest to the Sentencing Commission that a redraft of § 3D1.2(d) may be in order. One possibility would be to draft an application note relating to the list of specific offenses that “are to be grouped” under sub*185section (d).3 The new note could state that inclusion of the offenses at issue on this list raises a rebuttable presumption that grouping is appropriate so long as the -offense behavior is ongoing or continuous in nature, and so long as the purposes of the guidelines would be served by grouping the counts. By the terms of this presumption, grouping could be appropriate even in the absence of explicit language in the offense guidelines accounting for a continuing course of conduct. This would leave the grouping analysis to the judgment of the district court, rather than mandating its reliance on the formalities of the charging process.4
Another possibility would be to eliminate the “written to cover such behavior” clause entirely, and replace it with language requiring that the offense conduct underlying the potentially groupable counts be “closely intertwined.” In that case, subsection (d) would read:
... or if the offense behavior is ongoing or continuous in nature and the offense behavior underlying one of the counts is closely intertwined and represents substantially the same type of wrongful conduct as another of the counts.
Yet another possibility would be to replace the “written to cover” clause with a more general cross-reference to the policy goals outlined in the in the introductory comment to Chapter 3, Part D. In that case, the subsection could read:
... or if the offense behavior is ongoing and continuous in nature and the decision to group is necessary to prevent multiple punishment for substantially identical offense conduct.
I acknowledge that either type of amendment might generate new ambiguities in application. Such new ambiguities, however, would revolve around the actual facts of the ease and the policies of § 3D1.2, and not around the abstract language of the offense guidelines. I believe that such an change could help remove the charging process from the analysis, as the Commission apparently intended, and relocate sentencing discretion with the sentencing judge.
Honorable Edward R. Becker, United States Circuit Judge for the Third Circuit, assumed Chief Judge status on February 1, 1998.
. The problem with § 3D 1.2 highlighted by this case is also reflective of a larger, systemic shift caused by the adoption of the Guidelines, for that regime has, as many commentators have recognized, worked a massive transfer of discretionary power from the courts to the U.S. Attorneys’ offices. See Albert W,. Alschuler, The Failure of Sentencing Guidelines: A Plea for Less Aggregation, 58 U. Chi. L.Rev. 901, 926-28 (1991).
. Section 2C1.1(b)(1) provides for a base offense level enhancement if more than one bribe is involved in the offense. As the court notes, see Op. at 180, Rudolph was subject to this enhancement because "he took more than one bribe in exchange for INS templates.” If the PSI sale ■ had been charged as a bribe, it could also have been considered in this multiple-bribe analysis, rendering grouping of the offenses appropriate.
*183See also Op. at 176 (noting that had Rudolph’s two uncharged bribes been charged, they would have been groupable with the bribe charged in count one of the indictment).
. Section 3D 1.2(d) includes a list of offense guidelines, the offenses covered by which "are to be grouped under this subsection.” See § 3D 1.2(d). Our current construction of that subsection teaches that inclusion on this list does not mean that grouping is to be automatic, see United States v. Seligsohn, 981 F.2d 1418, 1425 (3d Cir.1992); see also United States v. Harper, 972 F.2d 321, 322 (11th Cir.1992) (cited in Selig-sohn) (grouping pursuant.to the subsection (d) "to be grouped" list not automatic because in some circumstances automatic grouping may detract from main purposes of § 3D 1.2).
. One can see this type of approach being applied in other cases that functionally construe § 3D 1.2(d). For example, the Eighth Circuit has held that a § 2F1.1 (fraud/counterfeiting) offense and a theft offense should be grouped under subsection (d) when they both arise out of the defendant’s theft and forgery of U.S. Treasury instruments. See United States v. Manuel, 912 F.2d 204, 206-07 (8th Cir.1990). The court concluded that:
Guideline section 3D 1.2(d) specifically lists Guidelines sections 2F1.1 and 2B1.1 as offenses that can be grouped together. U.S.S.G. § 3D 1.2(d). In light of the fact that these two counts both arose out of Manuel’s theft of United States Treasury instruments, we believe that the counts are sufficiently linked to merit being grouped together for purposes of section 3D 1.2(d).
Id. The Mantiel court, having found that the two offenses at issue are included among the "to be grouped” list of 3D1.2(d), did not find it necessary to probe further. Instead, it considered whether grouping on the facts of the case would advance the policies behind § 3D 1.2 by assessing whether the offenses were "sufficiently linked.”