[38] JOHN R. GIBSON, Circuit Judge, dissenting, with whom HEANEY, McMILLIAN, RICHARD S. ARNOLD, and MURPHY, Circuit Judges, join.

I dissent.

The Court today rejects circumstantial evidence of conspiracy and requires direct evidence to withstand summary judgment in an antitrust case. The court's requirement of direct evidence is contrary to Monsanto v. Spray-Rite Service Corp., 465 U.S. 752, 768 (1984), which only required "direct or circumstantial evidence that reasonably tends to prove . . . a conscious commitment to a common scheme designed to achieve an unlawful objective." (Emphasis added). Because conspirators cannot be relied upon either to confess or to preserve signed agreements memorializing their conspiracies, the court's requirement for direct evidence will substantially eliminate antitrust conspiracy as a ground for recovery in our circuit.

The potash industry is an oligopoly in which the producers ended a price war and raised prices dramatically. The question is whether the class has shown that the new prices resulted from an agreement among the producers to raise and stabilize prices, rather than from independent reactions to market conditions combined with actions of the United States and Canadian governments. I believe that the class has satisfied the existing standards for circumstantial proof that the prices resulted from collusion.

The market for potash in the United States is dominated by Canadian firms. In 1986, 84.3 per cent of the potash consumed in the United States was imported from Canada. The principal Canadian potash producers are defendants in this case: Potash Corporation of Saskatchewan Incorporated (PCS); Potash Corporation of America (PCA); IMC Fertilizer Group, Inc.; Kalium; Noranda Minerals, Inc.; and Cominco. These Canadian firms are allied in Canpotex, a cartel that exists to sell potash outside the United States.

Potash is a mineral that is an essential ingredient in fertilizer. Because potash is an essential ingredient, the demand for potash is "inelastic," meaning that people will continue to buy it even if the price goes up, and they will not buy much more, even if the price goes down. The effect of this inelastic demand is that low prices are bad for the producers because the low price does not result in greater sales, except insofar as one producer can take sales away from other producers. Conversely, producers benefit from high prices, because they can sell about as much potash and keep the extra money. Newcomers cannot enter the market without first gaining control of a potash mine, which is a significant entry barrier preventing new sellers from entering the market.

The biggest of the Canadian producers, PCS, was originally owned by the province of Saskatchewan and was run as a governmental company for the avowed purposes of providing jobs and promoting the local Saskatchewan economy. Unfortunately for the potash industry, due to a slump in agriculture, potash demand fell tremendously in the 1980's, resulting in oversupply. The effect of the oversupply was a potash price war, with prices bottoming in 1986 when PCS charged C$45.36 per ton FOB mine. The industry was in crisis. PCS alone lost $103 million in 1986. The president of PCS Sales wrote in an internal memorandum that the industry would not be able to end the price war without "joint action":

It is not possible for a single producer to affect [sic] a turn-around; however, joint action by a group of producers or governments could achieve this.

Given the competitive nature of the business, joint action in North America is not possible except through a vehicle such as Canpotex.

Canpotex by itself cannot achieve the objectives unless there is tacit approval and support on the part of other potash exporters. The danger inherent in multilateral decisions by Canpotex (or PCS Sales) is that the world will again see us as a residual supplier. . . . PCS Sales' past support of price with a view to achieve stability is proof of the fallacy of such attempts.

In fact, Noranda, Kalium, PCA, and PCS had each tried to increase prices unilaterally during 1986 and were forced to rescind the increases when the other producers undercut them.

In 1986, the province of Saskatchewan elected a government which promised to privatize PCS. In preparation for privatizing the company, PCS replaced its management with Charles Childers, CEO, and William Doyle, sales chief, who came to PCS from rival company IMC. Childers was quoted in a trade publication as saying that it was incumbent on PCS "to lead" in order to "straighten out our own company and hopefully give some strength to the potash industry as a whole."

Also in 1986, two American producers filed a complaint with the United States Department of Commerce alleging that the Canadian companies were dumping potash in the United States at less than fair value (which can mean below prices charged for exports to a third country, below domestic prices in the country where the product is produced, or below a reconstructed cost of production, see David G. Haglund and Alex von Bredow, U.S. Trade Barriers and Canadian Minerals: Copper, Potash and Uranium 65). The Department investigated the claim and in August 1987 issued a preliminary determination that the Canadian producers were dumping potash. The Department ordered the Canadian companies to post bonds on all exports to the United States, which would be payable to the United States as a duty if there were a final determination of dumping and injury to American producers. The amount of the bond varied for each firm according to the firm's "dumping margin," that is, the average amount by which the firm's United States sale price fell below the foreign market value in the cases examined by the Department. The dumping margins varied wildly, from 9.14 percent for IMC up to a crippling 85.2 percent for Noranda.

The Saskatchewan government responded to the United States action by adopting legislation which would give the province the power to limit and prorate production among Saskatchewan producers. See Haglund and von Bredow, supra, at 76. Within days of the introduction of the Saskatchewan legislation, on September 4, 1987, PCS announced that it would increase its prices by $35 per ton (from $58 to $93 per short ton for coarse grade FOB mine) to account for the bond expense. PCS's dumping margin was set at 51.9 per cent. PCS chose to raise its price only by $35, the amount necessary to pay duties on the industry average dumping margin of 36.62 per cent, rather than by the amount necessary to pay its own duty of 51.9 percent. It is undisputed that PCS chose the industry average figure in an attempt to pick a figure that other producers would follow. On September 11, 1987, the other Canadian producer defendants all announced either a $35 price increase or a new price of $93, rather than simply increasing the price by the amount necessary to cover their individual duty costs.

On January 8, 1988, the Canadian producers reached an agreement with the Department of Commerce, suspending the earlier order. The suspension agreement imposed a minimum price on Canadian producers selling in the United States market, but oddly, the minimum price chosen was still less than the "foreign market value": the Canadians were only forbidden from undercutting "foreign market value" by more than fifteen percent of the producer's dumping margin. See Haglund and von Bredow, supra, at 88. The literal terms of the suspension agreement made each producer's minimum price vary with its own foreign market value (which was arrived at by different methodologies for different firms) and with its own dumping margin; as the producers' expert William Barringer opined, the "uncertainties" in such a calculation make it impossible to predict with accuracy what price for any particular transaction would be in compliance with the agreement. However, correspondence from the Department of Commerce, which had responsibility for monitoring compliance with the suspension agreement, indicates that the Department used an industry average figure in assessing compliance.

As soon as the suspension agreement was in place, PCS again led the way in determining industry pricing, announcing it would rebate the earlier $35 surcharge, and on January 11, 1988, publishing its new price list at $86 per ton for granular grade, a net increase of $28 over the $58 it charged before the Department of Commerce imposed the bond requirement. The other Canadian producers matched PCS's increase within 11 days, and Kalium raised its price to $87. The Department of Commerce, in monitoring compliance with the agreement, noted that the agreement imposed an average floor price of $60.67, and that in the five months following the agreement the average price charged was $79.28.

The class's expert, Professor Gordon Rausser, opined and presented data purporting to show that the industry average prices for potash were higher than would be expected based on market factors during the 1988-1993 period (i.e., prices were "supra-competitive"). Perhaps more to the point, prices remained significantly higher than the minimum prices imposed by the suspension agreement from 1988 to 1992, with the exception of a short period in 1990 after PCS cut its prices drastically (and temporarily) for the avowed purpose of stabilizing prices within the industry. Only in January and February 1990, in the wake of this "market correction," did prices briefly dip below the suspension agreement price floor.

I.

This is a circumstantial evidence case based on parallel pricing, an example of the economic phenomenon called "conscious parallelism," in which competitors act alike. "[C]onscious parallelism is never meaningful by itself, but always assumes whatever significance it might have from additional facts." Donald F. Turner, The Definition of Agreement under the Sherman Act: Conscious Parallelism and Refusals to Deal, 75 Harv. L. Rev. 655, 658 (1962). The threshold question in deciding whether parallelism could be evidence of an agreement is whether the parallelism is "interdependent," meaning that the sellers would only take the course of action they took if they expected the others to behave the same way. Id. In this case interdependence is obvious; no producer would have been able to raise prices and keep them high unless the others did.

According to accepted economic theory, oligopolies are characterized by interdependent behavior because each seller is a big enough player to affect the market by price cuts; as a result, in raising or lowering prices or output, each seller must take into account his competitors' responses to his action. See Herbert Hovenkamp, Federal Antitrust Policy: The Law of Competition and its Practice 37 (1994); Turner, supra, at 665. In an oligopoly, price rises are not sustainable as long as one major seller continues to undercut the price leader and steal his customers. When one seller raises his price, no buyers will buy his product unless the other sellers follow the price leader and raise their prices. The other oligopolists know that if they keep their prices low, the brave price leader will simply cut his prices and the battle will resume. On the other hand, if they raise their prices in turn, all sellers will receive higher prices and end up with more money in their pockets. See Clamp-All Corp. v. Cast Iron Soil Pipe Inst., 851 F.2d 478, 484 (1st Cir. 1988) (Breyer, J.). The loser will be the consumer, who benefits from competition, not peaceful coexistence between suppliers.

Even though oligopoly pricing harms the consumer in the same way monopoly does, interdependent pricing that occurs with no actual agreement does not violate the Sherman Act, for the very good reason that we cannot order sellers to make their decisions without taking into account the reactions of their competitors. See Turner, supra, at 665-68. As then-Judge Breyer explained:

Courts have noted that the Sherman Act prohibits agreements, and they have almost uniformly held, at least in the pricing area, that such individual pricing decisions (even when each firm rests its own decision upon its belief that competitors will do the same) do not constitute an unlawful agreement under section 1 of the Sherman Act. That is not because such pricing is desirable (it is not), but because it is close to impossible to devise a judicially enforceable remedy for "interdependent" pricing. How does one order a firm to set its prices without regard to the likely reactions of its competitors?

Clamp-All, 851 F.2d at 484 (emphases in original) (citations omitted).

Although interdependent pricing tends to happen naturally in an oligopoly, there are good reasons for competitors to enter into an actual agreement to fix prices. First, successful price coordination requires accurate predictions about what other competitors will do; it is easier to predict what people mean to do if they tell you. In the absence of express agreements, oligopolists

must rely on uncertain and ambiguous signals to achieve concerted action. The signals are subject to misinterpretation and are a blunt and imprecise means of ensuring smooth cooperation, especially in the context of changing or unprecedented market circumstances. This anticompetitive minuet is most difficult to compose and to perform, even for a disciplined oligopoly.

Brooke Group Ltd. v. Brown Williamson Tobacco Corp., 509 U.S. 209, 227-28 (1993). Second, competitors may have different preferences on decisions such as pricing and therefore may not be willing just to follow a leader's decision; words (or word substitutes) may be necessary to negotiate a common course of action. Third, some oligopoly markets are more conducive than others to supra-competitive pricing. Market traits that make oligopoly pricing more likely to succeed include a standardized product and publicly announced prices, which make it easier for sellers to keep track of each others' prices. See Hovenkamp, supra, at 171. Actual agreement allows competitors to modify their market to facilitate collusion, particularly by setting up procedures for detecting and punishing price-cutting.

While the oligopoly market structure naturally facilitates supra-competitive pricing, that same market structure also makes cooperative arrangements unstable, for this reason: It is in the best interest of each individual competitor for his competitors to charge high prices, while he charges somewhat less when that will help him steal customers from his competitors. See George J. Stigler, A Theory of Oligopoly, 72 J. Pol. Econ. 44, 46 (1964) ("Let us assume that the collusion has been effected, and a price structure agreed upon. It is a well-established proposition that if any member of the agreement can secretly violate it, he will gain larger profits than by conforming to it."). The temptation to shade prices secretly is just as inherent in the oligopoly market structure as the temptation to collude to raise prices. See Hovenkamp, supra, at 140-41; Jonathan B. Baker, Two Sherman Act Section 1 Dilemmas: Parallel Pricing, the Oligopoly Problem, and Contemporary Economic Theory, 38 Antitrust Bulletin 143, 154 (1993). Of course, if the competitors know about the undercutting, they will match it. See Stigler, supra, at 46. Therefore, price-shading and secrecy must go hand in hand. While publicly announced prices discourage the sellers from cutting prices because they know that their price cuts will be matched, thus eliminating any competitive advantage, conversely, secretly negotiated discounts encourage price-cutting, since each seller hopes to steal customers without suffering retaliation from its competitors. See Hovenkamp, supra, at 141. As a result, a cartel can only succeed for any period of time if it has the ability to detect cheating and punish it effectively. See id.; Jacob Blinder Sons, Inc. v. Gerber Prods. Co. (In re Babyfood Antitrust Litig.), 166 F.3d 112, 137 (3d Cir. 1999) (without a mechanism to detect conspirator cheating, no conspiracy could long endure); JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775, 777 (7th Cir. 1999) (oligopoly market where cheating can be detected is "ripe" for collusion).

If the oligopolists agree, either tacitly or expressly, to coordinate price increases, they have committed a per se violation of section 1 of the Sherman Act. See United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 212-18 (1940) (agreement to fix prices per se illegal); American Tobacco Co. v. United States, 328 U.S. 781, 809-10 (1946) (agreement need not be express as long as there is unity of purpose or "common design and understanding"). From the outside, however, the conspirators' actions may look the same as innocent oligopoly pricing. Although parallel pricing evidence is consistent with illegal conduct, it is equally consistent with lawful conduct, and thus does not tend to exclude the possibility of independent action, as required by Monsanto Co. v. Spray-Rite Service Corp., 465 U.S. 752, 764 (1984). Therefore, parallel pricing in a concentrated market cannot make a submissible section one case, although it may set the groundwork for such a case. See State of Arizona v. Standard Oil Co. (In re Coordinated Pretrial Proceedings), 906 F.2d 432, 444 (9th Cir. 1990) ("We recognize that such interdependent pricing may often produce economic consequences that are comparable to those of classic cartels. Nonetheless, proof of such pricing, standing alone, is generally considered insufficient to establish a violation of the Sherman Act.").

In a rather primitive way, the "plus factors" test incorporates the economic principles outlined above as a way to distinguish between innocent interdependence and illegal conspiracy. Under this test, plaintiffs can establish a prima facie case of conspiracy by showing parallel prices together with "plus factors" that increase the likelihood that the parallel prices resulted from conspiracy. See, e.g., Wallace v. Bank of Bartlett, 55 F.3d 1166, 1168 (6th Cir. 1995); Petruzzi's IGA Supermarkets, Inc. v. Darling-Delaware Co., 998 F.2d 1224, 1232-33 (3d Cir. 1993); Todorov v. DCH Healthcare Auth., 921 F.2d 1438, 1456 n. 30 (11th Cir. 1991); Apex Oil Co. v. DiMauro, 822 F.2d 246, 253-54 (2d Cir. 1987).

We must, of course, take care to interpret the "plus factors" test in a way that is consistent with Monsanto. With Monsanto in mind, it is useful to distinguish between "plus factors" that establish a background making conspiracy likely and "plus factors" that tend to exclude the possibility that the defendants acted without agreement. For instance, "motive to conspire" and "high level of interfirm communications," are often cited as "plus factors" because they make conspiracy possible. See, e.g., Apex Oil, 822 F.2d at 254. Background facts showing a situation conducive to collusion do not tend to exclude the possibility of independent action, see id., but they nevertheless form an essential foundation for a circumstantial case. In Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574, 593-98 (1986), the Supreme Court held that a conspiracy case based on circumstantial evidence must be economically plausible. The background "plus factors" of market structure, motivation and opportunity play an important role in establishing such plausibility. Generally, these background "plus factors" are necessary but not sufficient to prove conspiracy.

On the other hand, acts that would be irrational or contrary to the defendant's economic interest if no conspiracy existed, but which would be rational if the alleged agreement existed, do tend to exclude the possibility of innocence. See Admiral Theatre Corp. v. Douglas Theatre Co., 585 F.2d 877, 884 (8th Cir. 1978); Re/Max Internat'l, Inc. v. Realty One, Inc., 173 F.3d 995, 1009-11 (6th Cir. 1999), petition for cert. filed, 68 U.S.L.W. 3138 (Aug. 17, 1999) (No. 99-294); City of Tuscaloosa v. Harcros Chems., Inc., 158 F.3d 548, 570-71 and n. 35 (11th Cir. 1998), cert. denied, 120 S.Ct. 3309 (1999).

A.

Of the "plus factors" that merely make conspiracy possible, such as motive and opportunity to conspire, the class has adduced abundant evidence. Within this category some "plus factors" are purely situational, involving no action on the part of the defendant, and some are volitional; while the former are important, the latter begin to make the required showing of collusion.

The purely situational factors in this case are the market structure and the crisis in the potash industry. The structure of the potash market was conducive to collusion, featuring an oligopoly, barriers to new sellers entering the market, inelastic demand, and a standardized product. See JTC Petroleum, 190 F.3d at 777 (market with these features is ripe for collusion). However, there was excess production capacity, which spurs competition, and a price war, which shows the producers had not been able to achieve a stable interdependent equilibrium. Individual attempts in 1986 by Noranda, Kalium, PCA, and PCS to initiate a price rise had failed. The producers were losing millions of dollars. The producers had good reason to wish for a truce.

The volitional background "plus factors" are also very strong in this case. At least one of the defendants actively considered the possibility of joint action, as is stated in the PCS "Corporate Plan" document dated September 25, 1986: "It is not possible for a single producer to affect [sic] a turn-around; however, joint action by a group of producers or governments could achieve this." Then there was a break in pattern, as the market went from price-war to profitability. See Turner, supra, at 672 ("Even in markets with few sellers, a fairly sudden change in pricing patterns is ground for suspicion.").

The class has introduced significant evidence of solicitations to enter a price-fixing agreement. Most, but not all, of the solicitations were by PCS. For instance, PCS freely complained to Kalium about Kalium's failure to adhere to pricing cut-offs. It was the custom in the industry to give lower prices at times of year when there was no immediate need for fertilizer, but to raise prices during high-use periods. Kalium published price lists announcing the pricing cut-off pattern, but in fact often shipped at the lower price after the cut-off date when it did not get orders filled before the cut-off date. PCS sales chief William Doyle repeatedly upbraided Kalium's vice president Robert Turner for shipping at the lower price after the cut-off date. Turner responded "something to the effect" that he would run his own business. Another time, Doyle called Turner and advised him that neither PCS, IMC, nor Cominco planned to accede to a certain customer's request to delay filling an order-that is, to ship at the old price after the cut-off. Turner answered that Kalium would try to ship by a certain date, as it had already said it would do in a letter to its customers. In the same vein, Doyle approached Turner about a certain bid and told Turner that Kalium's action was "wrong." John Ripperger, vice president of PCA, also testified that Doyle asked him if PCA was going to institute a price increase and not carry over product at the old price; Ripperger interpreted this question to mean that Doyle "would prefer that we don't make sales at the old price." Also, Doyle complained to Ripperger that PCA's pricing was undermining prices in Florida. Ripperger reported a conversation in which Gary Snyder of PCS asked a PCA salesman if he had sold at a certain price, and then said, "We [PCS] will take it [price] down and bury you [PCA] if that's what you want."

In 1988, after the sale of Kalium, Charles Childers, the CEO of PCS, called on Jay Proops, one of Kalium's new owners, armed with a chart showing that PCS was losing market share and that Kalium and other producers were gaining. Childers said Kalium was undercutting the price. Proops did some research and concluded that the chart had incorrect information and that Kalium was not undercutting. Therefore, Proops took no action in response to Childers's visit. In August 1990, Childers telephoned Joseph Sullivan, the other owner of Kalium. Childers told Sullivan that PCS's "price leadership was not working, despite major efforts" and that Childers "wanted to discuss this issue" with Sullivan. Sullivan declined to discuss prices.

Another time, a PCS employee took advantage of a trade meeting to apologize to Kalium's Turner about a low bid PCS had made by mistake. The PCS employee testified that he explained the mistake to Turner because he had "some concern that [the low price] may spread in the marketplace," and that he "was hopeful that it wouldn't go any darn further." Turner testified that Kalium matched the bid, but the reaction was "pretty much confined to that account. It did not go beyond that."

Though PCS made most of these overtures, on isolated occasions others did the same. Kalium's Turner called Ripperger of PCA to complain about a salesman who was cutting prices in Wisconsin. Similarly, Kip Williams of IMC complained to Ripperger about price-cutting in Florida.

Despite evidence that various defendants invited others to join in stabilizing prices, the class was not able to adduce direct evidence that the people on the receiving end of these solicitations accepted them and formed a deal. The evidence of solicitation is relevant, however, because it shows conspiratorial state of mind on the part of the solicitor and may also indicate that the solicitor was acting upon an earlier agreement. See 6 Phillip E. Areeda, Antitrust Law § 1419c (1986) ("Besides serving as direct evidence of a particular agreement, a solicitation might be circumstantial evidence of an ongoing conspiracy. Although no favorable response to the solicitation is shown, the solicitation itself might be the product of a prior agreement.") (footnote omitted).

B.

The stage was clearly set for conspiracy in this case. The question is whether the additional evidence tends to exclude the possibility that the producers acted independently. I believe that it does.

First, the class has produced evidence that the producers cooperated in disclosing prices they had charged on particular sales. The industry practice was that each producer published a price list stating its price, the dates for which that price would be available, and any discounts that the producer would extend. The price lists were widely distributed to customers and certainly were no secret. However, actual prices sometimes deviated from the lists. When Childers and Doyle came to PCS, a key aspect of their program to raise industry prices was to insist on the list price. Doyle stated in an industry publication: "When I first came on board in the spring of 1987, the first word I put out to our sales force was that the price list was our price, stick to that price and no bending. Anybody who bends was out of here." Despite published price lists with the high follow-the-leader price, the producers continued to undercut each other in privately negotiated deals. (This is what one would expect even from a cartel operating under an illegal agreement, see Hovenkamp, supra, at 143.) When word of the discounting got around to PCS, PCS executives, particularly sales chief Doyle, were quite active in contacting the discounter and asking for verification of the rumored price. Significantly, Doyle testified that he never made any such price verification calls before 1987. The number of these verification communications is difficult to pin down, but Doyle estimated he initiated or received three to four calls per year with PCA, five to six per year with IMC, three to four per year with Cominco, five to six total with Kalium, "a few" with NMPC, and one to two total with Noranda. Doyle was by no means the only person making such calls on behalf of PCS, and there is evidence that the other defendants called each other as well (except that there is no evidence of others calling Noranda).

These exchanges were often between high-level executives who were responsible for pricing decisions for their companies or who conveyed the price information to those who did set prices. For instance, Dale Massie, vice president of marketing for Cominco, testified that he had price verification communications with Doyle, head of sales at PCS. Massie testified that he made up the Cominco price lists, and the evidence shows that Doyle had a key role in determining PCS pricing policy. Charles Hoffman at IMC reported price information from Doyle to his superiors to inform them that "we would have to meet" PCS's price. Similarly, John Ripperger, vice president of PCA, had price verification discussions with Doyle, and Doyle said he had obtained price information from John Huber, Kalium's vice president of sales.

Price verification communications can either violate section 1 directly or they can be evidence of a violation. An agreement to exchange such communications can constitute an unreasonable restraint of trade under the rule of reason if the anticompetitive effect of the agreement outweighs its beneficial effects. See Penne v. Greater Minneapolis Area Board of Realtors, 604 F.2d 1143, 1148 (8th Cir. 1979); In re Coordinated Pretrial Proceedings, 906 F.2d at 447 n. 13. In price fixing cases, the exchange of sensitive price information can sometimes be circumstantial evidence of the existence of a per se violation. See Penne, 604 F.2d at 1149; In re Coordinated Pretrial Proceedings, 906 F.2d at 447 n. 13; In re Plywood Antitrust Litigation, 655 F.2d 627, 633-34 (5th Cir. Unit A 1981); Morton Salt Co v. United States, 235 F.2d 573, 577 (10th Cir. 1956) (competitors' exchange of price information "is a factor appropriately considered in determining the existence of a conspiracy"). It is this second theory that the class pursues in this case.

Again, acts that would be contrary to the actor's self-interest in the absence of a conspiracy, but which make economic sense as part of a conspiracy, provide the crucial type of "plus factor" evidence necessary to exclude the possibility of independent action. The class contends that "the price verification calls were inconsistent with the `pricing secrecy' sought by participants in oligopolistic industries because in such industries `each producer would like to secretly "shade" price[s], thereby gaining sales and avoiding retaliation.'" The class's argument finds support in the reasoning of United States v. United States Gypsum Co., 438 U.S. 422 (1978), which stated:

Price concessions by oligopolists generally yield competitive advantages only if secrecy can be maintained; when the terms of the concession are made publicly known, other competitors are likely to follow and any advantage to the initiator is lost in the process. Thus, if one seller offers a price concession for the purpose of winning over one of his competitor's customers, it is unlikely that the same seller will freely inform its competitor of the details of the concession so that it can be promptly matched and diffused.

Id. at 456 (citations omitted). Therefore, if there were no reciprocal agreement to share prices (and the producers certainly do not argue that there was), an individual seller who revealed to his competitors the amount of his privately negotiated discounts would have been shooting himself in the foot. On the other hand, if there were a cartel, it would be crucial for the cartel members to cooperate in telling each other about actual prices charged in order to prevent the sort of widespread discounting that would eventually sink the cartel.

Nor is there any legitimate business purpose which would make it desirable for the producers to reveal their pricing concessions notwithstanding the disadvantage of helping their competitors compete more effectively. These private communications between competitors had no redeeming effect of informing customers of prices, such as the advance announcements of price increases in Reserve Supply Corp. v. Owens-Corning Fiberglas Corp., 971 F.2d 37, 54 (7th Cir. 1992), or the advertisement of fees in Wallace v. Bank of Bartlett, 55 F.3d 1166, 1169 and n. 5 (6th Cir. 1995). Cf. Market Force, Inc. v. Wauwatosa Realty Co., 906 F.2d 1167, 1173 (7th Cir. 1990) (defendant broker announced intent to pay reduced commission to buyer's brokers; legitimate business reason was that other brokers needed to know in advance what commissions defendant was willing to pay). See also United States v. Citizens S. Nat'l Bank, 422 U.S. 86, 113-14 (1975) (correspondent bank program legitimate reason for "intimate and continuous cooperation and consultation on interest rates"). To the contrary, the prices stated here were discounts from the published price lists that reflected the prices the producers wanted to charge. The producers had no interest in publicizing these discounts to the market as a whole. Nor is there any evidence of special necessity for horizontal price communications, such as the customer fraud which justified the producers' practices in Cement Manufacturers Protective Ass'n v. United States, 268 U.S. 588, 595-96 (1925). The price communications in this case are more like those in In re Coordinated Pretrial Proceedings, 906 F.2d at 448, which served "little purpose" other than facilitating price coordination.

The Court today concludes that voluntarily revealing secret price-cutting to one's competitors is not probative of conspiracy, for three reasons, each of which is unsound.

First, the price verification communications involved completed sales, not future sales. The Court states: "Common sense dictates that a conspiracy to fix a price would involve one company communicating with another company before the price quotation to the customer." Supra at 10. This misconceives the purpose for which the price communications are being offered. The communications are not supposed to be direct evidence of a one-time mini-conspiracy to fix the price on one sale. Rather, they are circumstantial evidence of a type of behavior one would not expect in the absence of an agreement to cooperate. If no cartel was in place, each competitor would seek to benefit from high prices generally, while secretly shading prices when it would gain a customer without provoking retaliation. Confessing price-cutting when one needn't do so would only invite retaliation and guarantee that one's competitors could match the discounted price exactly next time. This is contrary to self-interest. On the other hand, if the producers were cooperating in a cartel, a necessary feature of their arrangement would be some way to determine who was discounting. Thus, confessing price-cutting to competitors makes no economic sense for independent actors, but makes perfect economic sense for cartel members. The Court has rejected circumstantial evidence of an agreement because it is not direct evidence.

The Court's second reason for dismissing the price verification evidence is that "[t]here is no evidence to support the inference that the verifications had an impact on price increases." Supra at 9. The class points to the price verifications as circumstantial evidence of a broader conspiracy. Parallel price increases are the starting point for the class's case, so that if the conspiracy is proved, effect on prices has been proved at the first step. Compare Court opinion, supra, at 10 ("Parallel pricing has been conceded") with In re Baby Food, 166 F.3d at 128-32 (parallel pricing not established). The Court argues that prices eventually went down, supra at 17, but this glosses over the fact that they first rose dramatically, then remained above both the forecasted price based on market factors and the suspension agreement price until 1992 (with the exception of the two-month dip caused by the PCS "market correction program"). If, to prove collusion, a plaintiff has to prove that there was no cheating, thus no downward pressure on prices, cartels will be quite safe from the Sherman Act. See generally Hovenkamp, supra, at 143, § 4.1a "The (Virtual) Universality of Cartel Cheating"; cf. United States v. SKW Metals Alloys, Inc., 195 F.3d 83, 89-90 (2d Cir. 1999) (in criminal price-fixing case involving widespread cheating by cartel members, conspiracy can "[affect] prices even when it falls short of achieving the conspirators' target price").

To support its proposition that there is nothing suspicious about oligopolists exchanging non-public price information, the Court relies on In re Baby Food Antitrust Litigation, where the Third Circuit stated: "No evidence . . . shows that any executive of any defendant exchanged price or market information with any other executive." 166 F.3d at 135. The court held that price discussions among low level employees did not show a conspiracy. Id. at 137. This reasoning implies that if high-level executives had been involved, it would have constituted evidence of conspiracy (if, indeed, the plaintiffs had been able to prove parallel pricing, which they did not in In re Baby Food, id. at 128-32). In our case there is a wealth of evidence that high level executives, who were in a position to respond to what they learned, were directly involved in exchanging secret price information. Citing In re Baby Food in a case with this kind of evidence vitiates the distinction on which the Third Circuit relied.

The Court ignores the Third Circuit's articulated rationale in In re Baby Food because the Court thinks the rationale is belied by facts in that case showing that high level executives used information gathered by low-level employees. Supra at 11. The Third Circuit apparently saw a crucial difference between gathering information to use to one's own advantage and giving out information for one's competitors to use to their advantage (and one's own detriment). However gladly the baby food executives used information relayed to them, the facts of In re Baby Food recited by the Third Circuit do not show that the executives were giving away their firms' secrets. The Third Circuit expressly stated that Anderson, a salesman who testified that he gathered information for his superiors, was not instructed to go around giving out advance pricing information about his own company. 166 F.3d at 126 ("Runk [Anderson's superior] never directed him to disseminate Heinz price information to competitors in the field"). In this case, it is the potash producers' pattern of giving their competitors valuable information that we identify as an act contrary to the producers' self-interest unless there was an agreement. Cf. 7 Up Bottling Co v. Varni Bros. (In re Citric Acid Litigation), 191 F.3d 1090, 1100 1110 (9th Cir. 1999) (existence of competitor's price list in defendant's files not probative because there is no evidence the competitor gave it to the defendant; defendant's disclosure of secret sales information to trade association was not an act contrary to self-interest because defendant was assured of confidentiality), petition for cert. filed, — U.S.L.W. — (U.S. Jan. 19, 2000) (No. 99-1218).

The Court's third reason for dismissing the price verifications is that the verifications were "sporadic." Supra at 10. The evidence indicates that the producers called each other when they had reason to think their competitors were cutting prices, and that they responded to each other's inquiries. The total number of such inquiries is difficult to set, but the defendants characterize it as "no more than several dozen" — surely more than a scintilla. Cf. Container Corp., 393 U.S. at 335 (liability where "all that was present was a request by each defendant of its competitor[s] for information as to the most recent price charged or quoted, whenever it needed such information" . . .; "[t]here was to be sure an infrequency and irregularity of price exchanges").

This "sporadic" argument seems to be directed to the quantum of proof, rather than the quality of it. In other words, it is an argument that more proof should exist, rather than an argument that the existing proof is not probative. If the plaintiff adduces evidence of the kind that tends to prove the existence of a conspiracy, I do not believe that Monsanto and Matsushita give a justification for rejecting it. Monsanto and Matsushita lay out a test for the kind of proof necessary in antitrust cases, not the quantity of it. Compare Monsanto, 465 U.S. at 764; and Matsushita, 475 U.S. at 587-88 ("antitrust law limits the range of permissible inferences from ambiguous evidence"; if claim makes no economic sense, plaintiff's evidence must be "more persuasive" than would otherwise be necessary) with Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 254-55 (1986) (where substantive law imposes a heightened standard of proof, as in libel cases, a higher quantum of proof is required to survive summary judgment). Cf. Merck-Medco Managed Care v. Rite Aid Corp., No. 98-2847, 1999 WL 691840, at *8 (4th Cir. Sept. 7, 1999) (unpublished per curiam) (holding that Monsanto-Matsushita inquiry requires higher quantum of probative evidence depending on relative strength of plaintiff and defendant's cases). There is no heightened "clear and convincing" standard of proof in civil antitrust conspiracy cases, requiring a greater quantum of proof than the ordinary "preponderance of the evidence" standard. See In re Brand Name Prescription Drugs Antitrust Litig., 186 F.3d 781, 787-88 (7th Cir. 1999), petition for certiorari filed, 68 U.S.L.W. 3327 (U.S. Nov. 5, 1999) (No 99-786). The plaintiff's evidence must amount to more than a scintilla, but the plaintiff does not have to outweigh the defendant's evidence item by item. See Rossi v. Standard Roofing, Inc., 156 F.3d 452, 466 (3d Cir. 1998).

The Ninth Circuit, in its recent decision of In re Citric Acid Litigation, dismissed evidence of price discussions as "sporadic," but it is not clear whether the court rejected the evidence qualitatively under Monsanto or because it was not quantitatively sufficient to create a material issue of fact. In Citric Acid the existence of a price-fixing conspiracy was conceded, but the question remained whether one defendant, Cargill, had been a party to the conspiracy. See 191 F.3d at 1093. Cargill received summary judgment despite testimony that a competitor had discussions regarding the bidding price of citric acid with someone from Cargill. The Ninth Circuit cited In re Baby Foods and dismissed the evidence as "sporadic price discussions with one individual at Cargill." Id. at 1105. The testimony in question was vague, with no specific evidence of what was said, and there were other qualitative reasons for rejecting it, including the fact that the timing of the alleged discussions was inconsistent with the plaintiff's theory of the case. See id. at 1104-05. The court also concluded the case by saying that there was no more than a scintilla of evidence that Cargill participated in the conspiracy. See id. at 1106. In re Citric Acid is therefore consistent with the ordinary summary judgment standard in cases requiring a preponderance of evidence, and it does not stand for the proposition that a heightened quantum of proof is necessary to survive summary judgment in antitrust cases.

The Court states in this case that the fact that there were "several dozen communications" among competitors is not "significant." Supra at 10. I would hold that evidence of several dozen communications of the type that tends to prove conspiracy creates a genuine issue of material fact.

In addition to the price verification practices, evidence concerning PCS's "market correction program" in December 1989 also tends to exclude the hypothesis of independent action. On December 18, 1989, PCS cut its prices by $18 a ton for five days. PCS's Carlos Smith stated that the purpose (and effect) of the program was to stabilize prices in the industry:

Q. [W]as it an attempt to stabilize prices?

A. Yes.

. . .

Q. And did it work?

A. It leveled them.

A high-level Kalium executive, John Huber, wrote the following notes: "Program was a market correction. Weren't trying to teach other people — only got tired of people who kept chipping away. Program was reasonable one — checked with people. . . . People started cheating. . . . We wanted to get their attention. Program to be short, very specific." (Emphasis added.) Huber said he did not recall to whom he had been talking when he made these notes, but the use of the phrase, "We wanted to get their attention," suggests he was taking dictation from someone at PCS.

The wording of Huber's notes implies that the "market correction" program was a way of disciplining producers who had breached an earlier agreement. In plain English, the use of the word "cheating" denotes the breach of an agreement or convention, not independent action. Without an agreement, price cutting would be called "competing," not "cheating."

Moreover, the Huber notes suggest PCS's action was not lonely price leadership, but rather that PCS "checked with people" before cutting prices. Apparently, PCS did not want to risk sparking another price war by letting other producers misunderstand the intent behind the "market correction program." These notes illustrate a situation in which smoke signals were just too ambiguous and dangerous to be trusted, so that competitors had to resort to explicit communications to coordinate prices. See Brooke Group Ltd. v. Brown Williamson Tobacco Corp., 509 U.S. 209, 227-28 (1993) ("anticompetitive minuet" difficult to compose and perform). The notes, taken together with the successful "market correction" program, tend to exclude the possibility of independent action. Cf. Petruzzi's IGA, 998 F.2d at 1233 (predatory pricing that looks like cartel disciplining cheaters is circumstantial evidence of conspiracy).

The class also points to another piece of evidence that tends to exclude the hypothesis of independent action. This is the Canpotex memorandum of Friday, January 8, 1988, which stated:

FYI Canadian potash producers have reached agreement with the United States Department of Commerce and all dumping action has been suspended for a minimum of 5 years. It is rumored that the USD per metric ton increase posted by Canadian producers in 1987 to cover possible tariff payments to the U.S. Govt will be refunded in full or part. In the meantime new price lists are being issued on Monday Jan. 11 at: Standard Grade USD 80.00; Coarse Grade USD 84.00; Granular Grade USD 86.00.

(emphasis added). Canpotex is the Canadian producers' cartel organized for sales outside the United States.

Again, the Court misconceives the import of this evidence, rejecting it as direct evidence of an attempt to reach an agreement, when the class offers it as circumstantial evidence of an agreement that already existed. The Court considers it crucial to establish who received the Canpotex memorandum, supra, at 13-14, apparently reasoning that if the memorandum was meant to negotiate an agreement, only people who got the memorandum could respond to it. Instead, the class offers this memorandum as circumstantial evidence tending to show that Canpotex knew on Friday, January 8, of an existing agreement to raise prices. The prophecy by Canpotex that its members would issue new "price lists" with particular prices does indeed tend to show the price increase was coordinated, because otherwise it would have been impossible to know in advance what the individual producers would do.

In sum, the class has adduced evidence of a market structure ripe for collusion, a sudden change from price war to supra-competitive pricing, price-fixing overtures from one competitor to another, voluntary disclosure of secret price concessions, an explicitly discussed cheater punishment program, and advance knowledge of other producers' price moves. Taken together, this list of "plus factors" adds up to evidence that satisfies the Monsanto standard.

II.

As I understand our previous cases, it is still necessary to take into account the producers' explanation of their conduct in order to ascertain whether their theory deprives the plaintiffs' case of its probative value. See Corner Pocket of Sioux Falls, Inc. v. Video Lottery Tech., Inc., 123 F.3d 1107, 1112 (8th Cir. 1997), cert. denied, 522 U.S. 1117 (1998); Lovett v. General Motors Corp., 998 F.2d 575, 580-81 (8th Cir. 1993). In this case, the producers' theory is that the price rises are explained entirely by the suspension agreement and the spectre of Saskatchewan prorationing legislation. In light of the evidence adduced by Professor Rausser and the Commerce Department correspondence indicating that the industry price far exceeded the price floors set by the suspension agreement, I cannot see that the producers' explanation deflates the class's price-fixing theory. I fully understand that the producers dispute Rausser's understanding of the suspension agreement price levels. Moreover, there is absolutely no dispute that prices had to rise above the price-war levels to comply with the suspension agreement. However, the producers make no attempt to identify a price floor required by the suspension agreement or to show that the producers actually set their prices, for instance the January 11 price of $86.00 for granular grade, by reference to the suspension agreement floor. Indeed, their expert William Barringer argued that it was actually impossible to ascertain whether a given price would satisfy the agreement. Instead, the producers' expert Andrew Rosenfield opined that the defendants needed only to set their prices "well above" suspension agreement floor prices. Under these experts' testimony, the suspension agreement did not dictate the actual prices charged; therefore, the existence of the suspension agreement does not explain away the facts supporting the class's theory that the actual prices were set by illegal collusion.

Not only were the January 1988 prices higher than the prices required by the suspension agreement, but there is evidence tending to prove that the producers felt free to dip below the suspension agreement prices when it served their purposes. Professor Rausser's price chart shows that during PCS's "market correction" program in January and February 1990, prices dipped below the suspension agreement floor. Deliberately taking prices below the suspension agreement floor to punish price-cutters is not a convincing sign of industry devotion to the suspension agreement, and the fact that it happened discredits the producers' argument that their prices were compelled by the suspension agreement.

The producers have not made a showing that the governmental intervention so explains their actual behavior as to take away the probative power of the class's case. This case should therefore proceed to trial. Accordingly, I dissent.