UNITED STATES of America, Plaintiff-Appellee, v. James P. HEFFERNAN, Defendant-Appellant, 43 F.3d 1144

POSNER, Chief Judge. James Heffernan, vice-president of a company that makes steel drums, conspired with executives of competing companies to sell the most common type of drum at identical prices to two large buyers. He was convicted of violating

section 1 of the Sherman Act, 15 U.S.C. § 1, and was sentenced to 24 months in prison. He would have received a shorter sentence had the judge not found that his offense was “bid-rigging” rather than simple price fixing. The federal sentencing guideline applicable to “Bid-Rigging, Price-Fixing, or Market-Allocation Agreements Among Competitors,” U.S.S.G. § 2R1.1 — the antitrust guideline, as it is known — imposes a one-level increase in the base offense level of a defendant whose offense involves the submission of “noncompetitive bids.” § 2R1.1(b)(1).

The commentary, oddly, does not repeat the term “noncompetitive bids.” Instead, like the title of the guideline, it uses, evidently as a synonym for “noncompetitive bids,” the term “bid rigging.” This is clearest in the background commentary, which states flatly: “the [Sentencing] Commission has specified a 1-level increase for bid-rigging.” The government, too, treats the terms as synonyms.

Its brief states: “bid rigging is simply the submission of any ‘subverted’ or ‘collusive,’ i.e., ‘noncompetitive’ bid”; “the identical bids submitted by Heffernan and his cohorts were surely non-competitive; indeed, they were plainly rigged.” And at oral argument, the government’s lawyer said, “I think the terms are meant synonymously, and there’s nothing in the guidelines that suggests otherwise.” The question presented by the appeal is whether Heffernan is indeed a bid rigger or, equivalently, a noncompetitive bidder, and this turns on what the term “bid rigging,” or the less common term “noncompetitive bids,” means.

The background commentary explains that the purpose behind the antitrust guideline is to specify the punishment for those antitrust violations about which there is “near universal agreement” that they “can cause serious economic harm.” These violations — the core per se offenses that no (well, very few) eeon-omists believe can be justified — are comprised of “restrictive agreements among competitors, such as horizontal price-fixing (including bid rigging) and horizontal market-allocation.” See generally Joseph C.

Gallo et al., “Criminal Penalties under the Sherman Act: A Study of Law and Economics,” 16 Research in Law and Economics 25 (1994). Heffernan argues that the term “bid rigging” and its synonym should be confined to bid rotation, what Application Note 6 calls the submission of “complementary bids”: for each job the competitors agree which of them shall be the low bidder, and the others submit higher bids to make sure the designated bidder wins. There was no bid rotation here; two purchasers solicited bids and the conspirators submitted identical bids.

The government argues that “bid rigging” includes all forms of collusion in a bidding process. The sentencing guideline and its commentary do not define “bid rigging.” For that matter they do not define “noncompetitive bids,” but as we have said they appear to treat these terms as synonyms. No case before today has had to decide what either term means in the guidelines. We affirmed the convictions of Heffernan and two of his coconspirators in United States v.

Rubin, 999 F.2d 194 (7th Cir.1993), but vacated the sentences because the defendants had been sentenced under the mail-fraud guideline rather than under the antitrust guideline, which we believed was error; so there was not yet any issue of adjusting a price-fixing sentence because of bid rigging. Before the guidelines were promulgated (and since, in eases that do not involve sentencing), the term “bid rigging,” though much bandied about in antitrust cases, did not denote a distinct offense. It was merely a descriptive term for a subset of price-fixing cases, so no one bothered with a careful definition.

Nevertheless, for what it is worth, we point out that the vast majority of cases in which the term has appeared have treated it as a synonym for bid rotation. See, e.g., United States v. Broce, 488 U.S. 563, 565, 577, 109 S.Ct. 757, 760, 766, 102 L.Ed.2d 927 (1989); United States v. Ward Baking Co., 376 U.S. 327, 328, 332, 84 S.Ct. 763, 765, 767, 11 L.Ed.2d 743 (1964); United States ex rel. Marcus v. Hess, 317 U.S. 537, 539 n. 1, 63 S.Ct. 379, 382 n. 1, 87 L.Ed. 443 (1943); United States v. Alex Janows & Co., 2 F.3d 716, 718 (7th Cir.1993); Moore v.

United States, 865 F.2d 149, 151 (7th Cir.1989); United States v. Walker, 653 F.2d 1343, 1345 (9th Cir.1981); United States v. All Star Industries, 962 F.2d 465, 470, 472 (5th Cir.1992); United States v. MMR Corp., 907 F.2d 489, 493 (5th Cir.1990); United States v. Pippin, 903 F.2d 1478, 1479 (11th Cir.1990); New York v. Hendrickson Bros., Inc., 840 F.2d 1065, 1084 (2d Cir.1988). We have mostly cited recent cases but could cite as many decided before the antitrust guideline was promulgated on November 1, 1987.

Consistently with the fact that the antitrust guideline equates “noncompetitive bids” to “bid rigging,” many of the cases we have cited, and others we could cite, use the terms interchangeably to mean bid rotation. See, e.g., United States v. Broce, supra, 488 U.S. at 565, 577, 109 S.Ct. at 760, 766; United States v. Ward Baking Co., supra, 376 U.S. at 328, 332, 84 S.Ct. at 765, 767; United States v. Alex Janows & Co., supra, 2 F.3d at 718; United States v. All Star Industries, supra, 962 F.2d at 473; United States v. MMR Corp., supra, 907 F.2d at 494; United States v. Pippin, supra, 903 F.2d at 1479; New York v.

Hendrickson Bros., Inc., supra, 840 F.2d at 1084; United States v. Evans & Associates Construction Co., 839 F.2d 656, 657, 661 (11th Cir.1988). The cases the government cites for a broader usage of the term “bid rigging” are inapt. The government’s favorite case, United States v. Portsmouth Paving Corp., 694 F.2d 312, 325 (4th Cir.1982), defines bid rigging as “any agreement between competitors pursuant to which contract offers are to be submitted to or withheld from a third party.” That is not remotely what happened in the present case. There was no agreement on who would bid, only on what the bid would be.

The court in Portsmouth did say that “a requirement that coconspirators agree to reciprocate by submitting complementary bids on future projects” is not part of the definition of bid rigging, id. at 325 (emphasis added). But the case was in fact a standard bid rotation case, see id. at 316 and n. 2, and the court thought the defendants’ violation more serious precisely because bids were rotated, which eliminated all competition rather than just price competition. Id. at 317.

The other cases that the government cites either quote Portsmouth but then go on to make clear that they understand bid rigging to mean bid rotation, as in United States v. Mobile Materials, Inc., 881 F.2d 866, 869, 871 (10th Cir.1989), and United States v. Reicher, 983 F.2d 168, 170 (10th Cir.1992), or simply do not indicate what they mean by bid rigging, as in Ramsay v. Vogel, 970 F.2d 471, 474 (8th Cir.1992). We have found only one case in which “bid rigging” is used to mean some other kind of interference with the integrity of a bidding system. It is Harkins Amusement Enterprises, Inc. v.

General Cinema Corp., 850 F.2d 477, 487 (9th Cir.1988), and it happens to be a post-guidelines ease, so it could not have influenced the Sentencing Commission in the drafting of the antitrust guideline. What is more, it was a case that involved an agreement to exclude a competitor from bidding — a little like bid rotation — not an agreement to submit identical bids.

Given that “bid rigging” appears to have had a reasonably settled meaning at the time the guidelines were promulgated, it is plausible (no stronger word is possible) that the draftsmen of the antitrust guideline meant to incorporate that meaning, in which event Heffernan would be entitled to be resen-tenced. (At the risk of becoming tedious, we repeat that the operative term in the guideline itself, as distinct from the title and commentary — “noncompetitive bids” — cannot be given a broader meaning.) But we should also consider the draftsmen’s purpose in singling out bid rigging for more severe punishment than other forms of price fixing, for that purpose might point to a broader meaning.

The background commentary explains that “volume of commerce is liable to be an understated measure in some bid-rigging eases. For this reason, and consistent with pre-guidelines practice, the [Sentencing] Commission has specified a 1-level increase for bid-rigging.” Application Note 6 further explains that “understatement of seriousness is especially likely in cases involving complementary bids.” It gives as an example a defendant who agrees not to submit a bid, or to submit an unreasonably high bid, on one occasion, in exchange for his being the designated low bidder on another occasion.

If he doesn’t get the second bid, he will not have any “volume of commerce”; and even if he does get the second bid, his volume of commerce will understate the volume affected by his participation in the bidding scheme. “The court should consider sentences near the top of the guideline range in such cases.” These two pieces of commentary (one in the background commentary, the other in Application Note 6) support the defendant’s argument that bid rigging means bid rotation.

The basic criterion that the guideline uses for determining the gravity of an antitrust violation is the volume of commerce attributable to the defendant, defined as “the volume of commerce done by him or his principal in goods or services .that were affected by the violation.” § 2Rl.l(b). If the defendant as part of a bid-rotation scheme does not submit a bid, and therefore does not make a sale, there may be, as we have pointed out, no commerce to attribute to him or (in this case) his employer, even though he has participated in a scheme that may have had a significant effect on commerce.

The simplest way to deal with this problem, and one that would be consistent with the usual treatment of conspiracy in the guidelines, U.S.S.G. § lB1.3(a)(l)(B); United States v. Blankenship, 970 F.2d 283, 288 (7th Cir.1992), would be to attribute the sales of each member of the conspiracy to all the others. The antitrust guideline does not do that. The only sales attributed to the conspirator are his own sales or, in the case of an individual fixing prices on behalf of his employer (or other principal), his employer’s sales.

Hence the need, as the draftsmen saw it, for a compensatory increase in sentencing when the defendant as part of the conspiracy has not submitted a bid (or at least a bid that any buyer would accept), hence has not made a sale, hence has no volume of commerce. Application Note 6 is expressly about bid rotation. It advises that in such cases the judge consider going to the top of the guideline range. But we know that the guideline requires a one-level increase in all bid-rigging cases.

If bid rigging means bid rotation, so that a one-level increase in the base offense level has been decreed in all bid-rotation cases, what sense does it make to advise the judge to go to the top of the guideline range in such cases on the ground that the guideline may understate the seriousness of the defendant’s conduct in a bid-rotation case? The one-level increase had been designed to avoid just such an understatement, if bid rigging means nothing more than bid rotation.

Another provision of the guideline, moreover, refers to “a bid-rigging case in’ which the organization submitted one or more complementary bids,” U.S.S.G. § 2Rl.l(d)(3), implying that some bid rigging does not or at least need not involve bid rotation.

Yet when bidding shenanigans do not take the form of bid rotation, the problem of understating the volume of commerce is no more serious than in any other price-fixing case — yet the danger of understatement is the only reason, other than “pre-guidelines practice,” not further specified, that the commentary gives for the one-level increase, and it is a reason applicable only to bid rotation — if there. For purposes of determining the amount of commerce, the guideline counts every sale just once, whether or not bids (rotated or otherwise) are involved.

If one bid rigger’s bid fails, another member of the ring gets the sale, and the sale is added to his volume of commerce. But the same is true if two ordinary price fixers make identical offers and one is accepted. So the guideline commentary itself does not dispel the mystery. But what about the pre-guidelines practice, to which the commentary refers? Many of the sentencing guidelines codify rather than alter pre-guide-lines sentencing practices. See United States Sentencing Commission, Guidelines Manual ch. 1, pt. A(3), p. 3 (Nov. 1994).

Unfortunately the parties have not told us anything about the pre-guidelines punishment of price fixing in bidding situations. Since the government has records of all such punishments, its failure to advise us concerning the pre-guidelines practice suggests, though very weakly in light of plausible alternative explanations, that the practice does not support its current position.

Our own research has turned up an empirical study which tentatively concluded that bid riggers were indeed punished more severely than other price fixers in the era before the guidelines, but the study does not define bid rigging and it suggests that the heavier sentences were attributable to the fact that government purchasers, the main users of sealed bidding procurement methods, are easy touches for price fixers. Mark A. Cohen and David T. Scheffman, “The Antitrust Sentencing Guideline: Is the Punishment Worth the Costs?” 27 Am.Cnm.L.Rev. 331, 344-47 (1989).

The most complete study of Sherman Act punishments, James C. Clabault & Michael K. Block, Sherman Act Indictments 1955-1980 (1981) (2 vols.), lists punishments for bid rigging separately but does not define the term, although a spot check of the eases indicates that the label usually refers to bid rotation and that when the case involves identical bids the authors classify it as simple price fixing. Another possibility, when the meaning of a guideline is in doubt, is to interpret it in accordance with the sentencing guidelines’ penal philosophy.

This is not easy to do in general, because the guidelines forswear the articulation of such a philosophy. See Guidelines Manual, supra, ch. 1, pt. A(3), pp. 3^4; Andrew von Hirsch, “Federal Sentencing Guidelines: Do They Provide Principled Guidance?” 27 Am.Crim.L.Rev. 367, 371 (1989). The statute authorizing the guidelines, however, makes reasonably clear that deterrence, incapacitation, retribution, and rehabilitation are the principal considerations in sentencing under the guidelines. 18 U.S.C. § 3553(a)(2); Mistretta v.

United States, 488 U.S. 361, 367, 109 S.Ct. 647, 652, 102 L.Ed.2d 714 (1989); United States v. Rivera, 994 F.2d 942, 949 (1st Cir.1993); United States v. Mogel, 956 F.2d 1555, 1558 (11th Cir.1992). We can ask which

interpretation of bid rigging, the broad or the narrow, would serve these goals better. No one supposes that antitrust defendants require rehabilitation to make them productive, law-abiding citizens upon their release from prison. Incapacitation — preventing this defendant from committing future crimes by keeping him in prison and thus out of circulation — does not appear to be an important element in the punishment of antitrust violators, either; their prison sentences are short by contemporary standards.

Considerations of retribution are normally thought to argue for proportioning punishment to the gravity of the crime. These considerations would be relevant here if the submission of identical bids were a graver antitrust violation than other forms of price fixing, and we shall consider that possibility. Deterrence is the goal most pertinent to the antitrust guideline.

Indeed, the background commentary says that “the controlling consideration underlying this guideline is general deterrence.” (“General” deterrence means deterrence of others besides the offender; “specific” deterrence means deterring this offender from repeating his offense.) Considerations of (general) deterrence argue for punishing more heavily those offenses that either are lucrative or are difficult to detect and punish, since both attributes go to increase the expected benefits of a crime and hence the punishment required to deter it.

Let us, therefore, merging considerations of retribution and of deterrence, consider whether bid rigging, in either the broad sense contended for by the government or the narrow sense (bid rotation) contended by for the defendant, is a graver, more lucrative, or more difficult to detect and punish offense than ordinary price fixing. Bid rigging in the broad as well as in the narrow sense has been an enormously common form of price fixing, or at least an enormous emphasis of the Department of Justice in enforcing the Sherman Act. Joseph C. Gallo, Joseph L. Craycraft, and Steven C.

Bush, “Guess Who Came to Dinner: An Empirical Study of Federal Antitrust Enforcement for the Period 1963-1984,” 2 Review of Industrial Organization 106, 126-27 (1985). A sealed-bid system of procurement makes it especially easy for price fixers to detect defectors from their conspiracy. Whether they agree to rotate bids or to submit identical bids, as soon as the bids are opened and the winner announced all .the bidders know whether the agreement has been violated, for they know who the winning bidder was and what his bid was.

The temptation of a member of a price-fixing conspiracy to cheat his fellows by shading the agreed price is very great, and is the bane of price' fixers, so anything that facilitates the prompt detection of a cheater firms up the conspiracy. In addition, sealed bidding tends to be used disproportionately by government agencies (because of traditional concerns about bribery of public employees), which, lacking as they do market incentives to minimize their-costs, are often especially vulnerable to overcharging by their suppliers.

Sealed bidding also tends to be used in the purchase of inputs of a simple, standardized character, such as the steel drums involved in this case, where the costs of individual negotiations with potential suppliers may exceed the benefits. These are also the products that it is easiest to fix the prices of, because there are few other dimensions on which the conspirators would also have to agree in order to eliminate competition among themselves.

Finally, when the buyer is asking suppliers to bid on a specified quantity, the normal effect of monopoly or cartel pricing in reducing the quantity of the product demanded is unlikely to operate unless the bids are really exorbitant; so the cartel profits are likely to be greater because earned on a larger output. But except for the point about the simple product, none of the arguments for punishing price fixing of bids more seriously than other price fixing has any application to this case, which for all we know is representative of a great number of cases involving the submission of identical bids.

The purchasers were not government agencies, .but large private firms presumably well able to take care of themselves. They were not asking for bids on a fixed quantity of drums — they asked for bids on a number of different quantities. They did not employ sealed bids, and in fact one of the purchasers was accustomed to asking the high bidders to lower their prices. This kind of “bid rigging” is indistinguishable from ordinary price fixing, in which competitors get together and agree to sell at a uniform price. Which is all that happened here.

To punish Heffernan more heavily than an ordinary price fixer merely because his customers asked for “bids” rather than “offers” would be irrational. This irrational result is avoided by interpreting “bid rigging” (equivalently “noncompetitive bids”), consistently with the cases and the pre-guidelines sentencing practice, as meaning bid rotation. Since Heffernan did not engage in bid rigging in this sense, he should not have received the one-level increase in his base offense level. That is our best guess as to the meaning of the antitrust guideline.

But we invite the Sentencing Commission to rewrite the guideline; its treatment of bidding is a muddle. Heffernan’s sentence is vacated and the case is remanded with instructions to resen-tence him in conformity with this opinion. So ORDERED.