concurring:

Judge Wallace concludes that the trial court’s instructions on the restraint of trade claim were erroneous and remands for a new trial. In its briefs before this court, General Portland attempted to defend the challenged instructions. But it spent more time arguing that any error was harmless. General Portland claims that French offered no evidence of anticom-petitive effect in a relevant market. Since proof of anticompetitive effect is a necessary element of a rule of reason case, General Portland argues that French’s claim must fail regardless of the trial court’s instructions to the jury.

It is well established that proof of anti-competitive effect is essential to a rule of reason case. See, e.g., Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 487 , 97 S.Ct. 690, 696 , 50 L.Ed.2d 701 (1977). It is also beyond dispute that General Portland has preserved its objections on this ground by moving for a directed verdict at trial and by filing a cross-appeal from the denial of its motion. If General Portland’s argument that French did not introduce sufficient evidence of anticompet-itive effect were correct, we would be obliged to affirm. Thus, although he does not explicitly address the issue, Judge Wallace’s silence can only be interpreted as a rejection of General Portland’s argument. In the interests of fairness to General Portland, the reasons for this holding should be made explicit.

French argues that the conspiracy between Conrock and General Portland had an anticompetitive effect in the West Los Angeles concrete market. General Portland is correct in its assertion that French did not attempt to define the relevant market in any of the traditional ways. It did not offer direct economic evidence of the cross-elasticity of demand between concrete and other construction materials, or of the sensitivity of the price of concrete in West Los Angeles to changes in the price of concrete produced elsewhere. It did not have an industry expert take the stand and testify about these matters. It provided only brief and sketchy evidence of topography, traffic patterns and transportation costs which may tend to insulate firms in West Los Angeles from outside competitive pressures.

*1403 However, French did offer evidence that it was a producer of concrete in West Los Angeles, that the other producers of concrete in West Los Angeles had formed and carried out a sustained price-fixing conspiracy, that there were no producers of concrete in the West Los Angeles area other than French who were not involved in the price-fixing conspiracy, and that the Con-rock-General Portland conspiracy had succeeded in eliminating French. French argues that this evidence is extremely probative on the issue of anticompetitive effect.

As a purely logical matter, French is unquestionably correct. A price-fixing conspiracy confined to manufacturers of concrete would not have been able to succeed if concrete were not a distinct product market: when the cartel attempted to raise prices, customers would simply switch to sand, brick, gravel or some other construction material. Similarly, a price-fixing conspiracy confined to firms in West Los An-geles would not have succeeded if West Los Angeles were not a distinct geographic market: when the cartel attempted to raise prices, customers would simply take their business to East Los Angeles or San Diego or Phoenix. If a group of firms is able to fix prices, it is because their customers have nowhere else to turn. Every price-fixing conspiracy thus identifies directly, in a real world context, a group of firm which is insulated from outside competitive pressures. That is precisely what conventional market definition evidence attempts to identify artificially, by the collection and interpretation of economic data regarding the relationship between various demand curves, by common sense assumptions about the interchangeability of similar products, and the like.

If French’s evidence is believed, the only issue is whether the elimination of the only firm in the relevant market not involved in a price-fixing conspiracy had a likely adverse effect on price or output. That question, of course, is virtually circular. So long as French was around, at least some concrete customers in West Los Angeles— those served by French — could hope to pay prices set by market forces rather than by a cartel. When French was eliminated, that partial protection was lost. In addition, the cartel need no longer set its prices low enough to prevent the loss of additional customers to French. Thus, all concrete customers in West Los Angeles will very likely be worse off as a result of the Con-rock-General Portland conspiracy. Although French’s evidence is not in the familiar forms, it has enough persuasive force to raise a triable issue.

There are situations in which antitrust courts will quite properly reject even extremely persuasive evidence for reasons of uniformity, predictability, ease of administration, or the like. In Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227 (1st Cir.1983), for example, the issue was whether pricing below marginal cost but above average cost constitutes predatory or exclusionary conduct under section 2 of the Sherman Act. The First Circuit recognized that a firm maximizes profits in the short run by pricing at the level of marginal cost. Prices below that level raise a strong inference of predation. But while the Sherman Act condemns acts of predation, it is ultimately designed to promote lower prices. Although prices below marginal cost theoretically fall on the wrong side of the line, there is a danger in trying to draw the line too precisely. Courts are fallible. If they err in their quest for theoretical precision, they end up punishing the very conduct the Sherman Act was designed to foster. In Barry Wright, the First Circuit held that the risk of error was so substantial that a court should not concern itself with predatory pricing unless prices fall below average cost and the firm is actually losing money in the short run. Actual losses are a manageable standard which courts can apply with little risk of error.

Under similar reasoning, it might be argued that French’s evidence, although extremely probative as a purely logical matter, should not be enough to get to the jury. The most familiar and straightforward method of proving a relevant market is to offer direct evidence about product demand, interchangeability of use, transportation costs, etc. See, e.g., United States v. E.I. Du Pont de Nemours & Co., 351 U.S. 377, 395 , 76 S.Ct. 994, 1007 , 100 *1404 L.Ed. 1264 (1977). These are the methods courts are most comfortable administering, and they presumably contain the least risk of error. While there may be exceptional cases where a relevant market can be proven without relying on the traditional methods, recognizing the exception might invite error in a less clear case, and in general might not be worth the trouble.

Our holding, today is most significant in rejecting this argument. There is evidence in the record of real world pricing behavior which indicates that West Los Angeles concrete manufacturers occupy a distinct market. Requiring plaintiff to assemble economic data or other evidence verifying this observation would invite needless waste; and defining a relevant market through traditional methods is notoriously complicated, expensive and time-consuming.

More to the point, evidence of an overlapping price-fixing conspiracy may be more probative and less susceptible to error than the traditional methods of proving anticompetitive effect. The concept of a single “relevant market” is itself an abstraction. Competition is a matter of degree. The demand for Pepsi, for example, may be extremely sensitive to changes in the price of other colas, somewhat sensitive to changes in the price of root beer, and still less sensitive to changes in the price of mineral water. It would be an oversimplification to draw the line at a single point along this continuum. A conclusion that the relevant product market is soft drinks, for example, will be both overinclusive, to the extent that Pepsi and root beer are not perfect substitutes, and underinclusive, to the extent that some Pepsi customers would just as soon drink mineral water. Defining a relevant geographic market involves similar uncertainties. See, e.g., United States v. Philadelphia National Bank, 374 U.S. 321, 360, n. 37 , 83 S.Ct. 1715, 1735, n. 37 , 10 L.Ed.2d 915 (1963) (“fuzziness would seem inherent in any attempt to delineate the relevant geographical market”). Even perfect knowledge of the relevant cross-elasticities of demand cannot ordinarily resolve questions about market structure with anything approaching scientific certainty. A certain amount of error is built into the process.

Moreover, the market definition approach is only an indirect and attenuated way of measuring anticompetitive effect. The ultimate issue in a rule of reason case is whether a challenged practice will produce adverse effects on price or output. See, e.g., FTC v. Indiana Federation of Dentists, 476 U.S. 447, 460 , 106 S.Ct. 2009, 2018 , 90 L.Ed.2d 445 (1986); National Society of Professional Engineers v. United States, 435 U.S. 679, 690, n. 15 , 98 S.Ct. 1355, 1364, n. 15 , 55 L.Ed.2d 637 (1977). The only direct way to answer that question is to introduce evidence of actual price increases or reductions in output after the challenged practice. But even if a plaintiff is lucky enough to gather such evidence, he will face the momentous task of proving that the observed price or output effects were not attributable to any one of an infinite number of independent causes: exhaustion of raw materials, increases in labor costs, increases in the price of substitute goods, tax hikes, etc. Situations will arise where a plaintiff is able to meet this burden. See, e.g., Indiana Federation of Dentists, 476 U.S. at 460-61 , 106 S.Ct. at 2018-19 . But doubtless those occasions will be rare.

Courts have developed the market definition approach to proving anticompetitive effect as a way around these difficulties. If, as a practical matter, we cannot show anticompetitive effect from a before-and-after comparison, we might at least be able to predict what the consequences of a given practice will be by examining the context in which the practice occurred. Where the challenged practice results in the elimination of a single firm, the market definition approach predicts likely competitive impact by asking whether the market from which the firm was eliminated was strong enough to bear the loss: Are firms producing similar products effective competitors for the business of the eliminated firm and therefore able to take up the slack (relevant product market evidence)? Can firms in surrounding localities compete in the eliminated firm’s territory and therefore take up the slack (relevant geographic market evidence)? Was the eliminated firm fairly small relative to firms not excluded *1405 through the market definition evidence, so that the survivors can expand and take up the slack; are barriers to entry insubstantial, so that new firms can enter the market and take up the slack; do conditions within the affected industry make it difficult to exchange price information and enforce collusive agreements, so that the remaining firms will find it hard to aggregate power (evidence of impact within the relevant market)? In short, the “market definition” method tends to prove harm to competition by demonstrating that the ordinary corrective factors of a healthy free market economy are weakened or absent.

The crucial point is that neither market definition nor the traditional methods of market definition are essential to proving anticompetitive effect. See Indiana Federation of Dentists, 476 U.S. at 460-61 , 106 S.Ct. at 2018-19 ; Oahu Gas Services, Inc. v. Pacific Resources Inc., 829 F.2d 1471, 1474 (9th Cir.1987); Southern Pacific Communications Co. v. AT & T, 740 F.2d 1011, 1020 (D.C.Cir.1984). Market definition is simply one way of addressing an extremely complicated issue. It is usually the best we can do, but it hardly eliminates the possibility of error.

The evidence in this case is more direct than the type usually relied on by antitrust courts. The jury need not make what are necessarily somewhat arbitrary judgments about which group of firms provide a meaningful context for estimating the likely effects of the Conrock conspiracy. The jury need not evaluate complicated evidence about the conditions within that relevant market in order to ascertain whether the elimination of French will leave consumers susceptible to the exercise of market power. French introduced evidence that West Los Angeles concrete manufacturers were already exercising power over price. Whether or not the cartel faced some competitive pressures from other products or other areas, it certainly faced competition from an independent firm producing the exact same product in the exact same area. The elimination of that firm necessarily led to an increase in the cartel’s power. Of course, a jury must ultimately decide whether French’s evidence is credible. But submitting French’s claim to the jury does not entail a substantial risk of error.

Nor is there any reason to believe that our holding will be misapplied in future cases. Today’s facts are somewhat exceptional. The members of the Conrock conspiracy were unfortunate enough to have directed their actions toward a market which was already dominated by a separate price-fixing conspiracy; because they produced the same product in the same area, French and the members of the price-fixing conspiracy were part of the same market. We have found no decision which has considered this precise situation, but previous cases have addressed analogous issues.

In Syufy Enterprises v. American Multicinema, Inc., 793 F.2d 990 (9th Cir.1986), the plaintiff argued that the relevant market was limited to “anticipated top-grossing films” in the San Jose area. Plaintiff “did not present direct evidence going to the ‘cross-elasticity of demand’ between this limited class of pictures and other films.” 793 F.2d at 995 . However, it did offer evidence that anticipated top-grossing films received national advertising support, were booked in first class theatres, and in general were treated differently than other films. We held that this evidence was sufficient to define the relevant market. Since interested market participants treated and regarded anticipated top-grossing films as occupying a separate market, a jury could conclude that the films occupied a separate market. Id. at 994-95 . Similarly, in Ralph C. Wilson Industries v. Chronicle Broadcast, 794 F.2d 1359 , 1363-64 (9th Cir.1986), a case presenting a challenge to exclusive television licensing practices, we based our determination of the relevant market primarily on classifications used by the two national television ratings services.

The evidence introduced by French falls within the rationale of these cases. Evidence that West Los Angeles concrete manufacturers entered into a price-fixing conspiracy indicates that interested market participants — those in the best position to know — believed that they were insulated from outside competitive pressures. The *1406 West Los Angeles concrete manufacturers were, if French’s evidence is believed, able to sustain a price-fixing conspiracy. As noted above, firms cannot maintain a price-fixing conspiracy unless they are in fact insulated from outside competition. Under Syufy and Ralph C. Wilson, evidence of price-fixing is sufficient to define the relevant market.

On remand the district court can determine whether French may bolster its proof of anticompetitive effect by introducing any of the traditional types of market definition evidence. In reversing and remanding, we merely recognize that French raised a genuine issue of material fact as to anticompetitive effect even though it failed to introduce traditional types of market definition evidence.