I. Background

Centerprise is a British corporation that uses DRAM in the manufacture of its computers. DRAM is a common type of memory chip that is sold around the world. According to Cen- terprise, DRAM is “a readily transportable commodity prod- uct with multiple firms offering essentially identical parts.” Centerprise purchased DRAM outside of the United States from the defendants, various memory companies.

Centerprise brought this antitrust class action in May 2005 on behalf of itself and all others similarly situated, pursuant to §§ 4(a), 12 and 16 of the Clayton Act, 15 U.S.C. §§ 15, 22 and 26, seeking injunctive relief and damages, premised on defendants’ alleged violations of federal antitrust laws, including § 1 of the Sherman Act.2 Centerprise alleged that the defendants engaged in a global conspiracy to fix DRAM prices, raising the price of DRAM to customers in both the United States and foreign countries. Specifically, Centerprise asserted that the domestic effect of the defendants’ anticom- petitive conduct — higher DRAM prices in the United States — gave rise to its foreign injury of having to pay higher DRAM prices abroad because the defendants could not have raised prices worldwide and maintained their global price- fixing arrangement without fixing the DRAM prices in the United States.

The district court dismissed the complaint with prejudice for lack of subject matter jurisdiction under the FTAIA. Rely- 2 Centerprise defined the class as “[a]ll individuals and entities located outside of the United States who, during the period from approximately July 1, 1999 through at least June 20, 2002 (the ‘Class Period’), purchased DRAM directly from defendants, any subsidiaries or affiliates thereof.” 10650 IN RE DYNAMIC RANDOM ACCESS MEMORY ing on the Supreme Court’s decision in F. Hoffmann-La Roche Ltd. v. Empagran S.A., 542 U.S. 155 (2004) (“Empagran I”), and the D.C. Circuit’s decision in that case on remand, the district court held that Centerprise had not met the jurisdictional prerequisites under the FTAIA because it had not sufficiently alleged that its foreign injury was directly linked to the domestic effect of higher U.S. prices for DRAM. The district court also denied Centerprise leave to amend its complaint as futile because its proposed amendments did not substantively change its theory of recovery. Centerprise timely appealed.

II. Discussion

A. Legal Standards

We review de novo the district court’s dismissal for lack of subject matter jurisdiction. See United States v. LSL Biotech- nologies, 379 F.3d 672, 677 (9th Cir. 2004). The party assert- ing jurisdiction bears the burden of establishing subject matter jurisdiction on a motion to dismiss for lack of subject matter jurisdiction. See Kokkonen v. Guardian Life Ins. Co. of Am., 511 U.S. 375, 377 (1994); see also Stock W., Inc. v. Confeder- ated Tribes of the Colville Reservation, 873 F.2d 1221, 1225 (9th Cir. 1989). Dismissal for lack of subject matter jurisdic- tion is appropriate if the complaint, considered in its entirety, on its face fails to allege facts sufficient to establish subject matter jurisdiction. See Love v. United States, 915 F.2d 1242, 1245 (9th Cir. 1990).

B. Subject Matter Jurisdiction

[1] In 1982, Congress responded to concerns regarding the scope of the broad jurisdictional language in the Sherman Act by enacting the FTAIA.3 See Phillip E. Areeda & Herbert 3 Section 1 of the Sherman Act prohibits “[e]very contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations . . .” 15

U.S.C. § 1.

IN RE DYNAMIC RANDOM ACCESS MEMORY 10651

Hovenkamp, Antitrust Law ¶ 272i, pp. 286-87 (3d ed. 2006) (hereinafter “Areeda & Hovenkamp”). The FTAIA amends the Sherman Act and “excludes from [its] reach much anti- competitive conduct that causes only foreign injury.” Empa- gran I, 542 U.S. at 158. It does this by establishing a general rule that the Sherman Act “shall not apply to conduct involv- ing trade or commerce . . . with foreign nations.” § 6a. It then provides an exception to this general rule, making the Sher- man Act applicable if foreign conduct “(1) has a ‘direct, sub- stantial, and reasonably foreseeable effect’ on domestic commerce, and (2) ‘such effect gives rise to a [Sherman Act] claim.’ ” Empagran I, 542 U.S. at 159 (quoting § 6a) (alter- ation in Empagran I).4 This exception is known as the “do- mestic injury exception” of the FTAIA. The Supreme Court has described the FTAIA’s language as:

initially lay[ing] down a general rule placing all (nonimport) activity involving foreign commerce outside the Sherman Act’s reach. It then brings such conduct back within the Sherman Act’s reach pro- vided that the conduct both (1) sufficiently affects American commerce, i.e., it has a “direct, substan- tial, and reasonably foreseeable effect” on American 4 The FTAIA states: Sections 1 to 7 of [the Sherman Act] shall not apply to conduct involving trade or commerce (other than import trade or import commerce) with foreign nations unless — (1) such conduct has a direct, substantial, and reasonably fore- seeable effect — (A) on trade or commerce which is not trade or commerce with foreign nations, or on import trade or import commerce with foreign nations; or (B) on export trade or export commerce with foreign nations, of a person engaged in such trade or commerce in the United States; and (2) such effect gives rise to a claim under the provisions of sec- tions 1 to 7 of this title, other than this section. § 6a. 10652 IN RE DYNAMIC RANDOM ACCESS MEMORY domestic, import or (certain) export commerce, and (2) has an effect of a kind that antitrust law considers harmful, i.e., the “effect” must “giv[e] rise to a [Sherman Act] claim.”

Id. at 162.

[2] The FTAIA thus clarifies that U.S. antitrust laws con- cern the protection of “American consumers and American exporters, not foreign consumers or producers.” Areeda & Hovenkamp at ¶ 272i, pp. 287. For the Sherman Act to apply, the effect on U.S. commerce or American interests engaged in foreign commerce must be direct, substantial and reason- ably foreseeable — not minor impacts — and it must “giv[e] rise” to the antitrust claim. See id.

[3] In dismissing Centerprise’s action, the district court concluded that Centerprise had sufficiently alleged that defen- dants’ conduct had a “direct, substantial, and reasonably fore- seeable” effect on U.S. domestic commerce, the first prong of the domestic injury exception, but did not sufficiently allege the second prong, that such U.S. domestic effect “g[ave] rise to” Centerprise’s foreign injury under § 6a of the FTAIA. Only the district court’s conclusion with respect to the second prong of the domestic injury exception is at issue in this appeal.5

[4] The controlling precedent on the FTAIA domestic injury exception is the Supreme Court’s opinion in Empagran

I. There the Supreme Court addressed the exception in the

context of an antitrust class action brought by foreign pur- chasers of vitamins who alleged a global price-fixing conspir- acy that led to increased prices in the United States, and 5 We focus on whether the exception to the FTAIA’s general rule excluding application of the Sherman Act is available because it is undis- puted that the alleged price-fixing conduct falls within the FTAIA’s gen- eral rule because the price-fixing activity constitutes “conduct involving trade or commerce . . . with foreign nations.” § 6a.

IN RE DYNAMIC RANDOM ACCESS MEMORY 10653

independently led to increased prices for vitamins in other countries. 542 U.S. 155. The Court stated the issue as whether the Sherman Act reaches “anti-competitive price-fixing activ- ity that is in significant part foreign, that causes some domes- tic antitrust injury, and that independently causes separate foreign injury.” Id. at 158.

[5] After considering principles of comity and the history of the Sherman Act and the FTAIA, the Court concluded that “Congress would not have intended the FTAIA’s exception to bring independently caused foreign injury within the Sherman Act’s reach.” See id. at 173. Thus the foreign purchasers who bought vitamins outside of the United States could not bring a claim under the Sherman Act “where [their] claim rests solely on the independent foreign harm.” Id. at 159. The Court remanded the case to the D.C. Circuit to consider the plaintiffs’ alternative argument that the foreign injury was linked to the domestic effects, and not independent of them. Id. at 175. The Court left open the question of the standard to apply in determining whether the plaintiffs alleged a sufficient link between the U.S. effect and their foreign injury so as to establish subject matter jurisdiction under the Sherman Act.6 6 The Areeda & Hovenkamp treatise comments on this open question as follows: Clearly [the Supreme Court] could not have meant that the mere fact that this was a conspiracy involving both sellers and buyers all over the world was sufficient. To be sure, in such a conspiracy the injury to any particular buyer is linked to the injury suffered by other buyers or else the conspiracy would not work. . . . But the forbearance of one cartel member from cutting price or shipping into another cartel member’s territory is necessary to the functioning of every cartel. To interpret “linkage” of foreign and domestic injury this broadly would have undermined the entirety of the Court’s opinion, which unambiguously held that foreign plaintiffs injured by a conspiracy that also injured American pur- chasers could not sue under the Sherman Act. Areeda & Hovenkamp ¶ 272i5, pp. 317-18. 10654 IN RE DYNAMIC RANDOM ACCESS MEMORY [6] On remand, the D.C. Circuit held that the “gives rise to” language of § 6a of the FTAIA domestic injury exception requires a direct or proximate causal relationship between the domestic effect and the foreign injury, not merely a “but for” relationship. See Empagran S.A. v. F. Hoffmann-Laroche, Ltd., 417 F.3d 1267, 1271 (D.C. Cir. 2005) (“Empagran II”). The plaintiffs’ asserted theory on remand was that the domes- tic effect of the anticompetitive conduct — higher U.S. vita- min prices — gave rise to their foreign injury of higher vitamin prices abroad because the defendants could not have maintained their global price-fixing arrangement without fix- ing the prices in the United States as well. The court con- cluded that under this theory the domestic effect of the conspiracy was only a “but for” cause of the plaintiffs’ inju- ries, so the requirements of the FTAIA exception were not met and the Sherman Act was therefore inapplicable. The Eighth Circuit recently joined the D.C. Circuit in interpreting the FTAIA’s domestic injury exception to require proximate cause. See In re Monosodium Glutamate Antitrust Litig., 477 F.3d 535, 539 (8th Cir. 2007).

We have not had occasion since Empagran I to illuminate the standard that applies in determining whether a domestic effect “gives rise to” a foreign injury, where the plaintiff alleges its foreign injury was linked to the domestic effect of the defendants’ anticompetitive conduct. Like the plaintiffs in Empagran II, Centerprise contends the FTAIA’s domestic injury exception requires only “but for” causation and that its claim meets this standard. Alternatively, Centerprise contends that its claim satisfies even a proximate cause standard.

[7] Like the D.C. Circuit and the Eighth Circuit, we con- clude that “but for” causation cannot suffice for the FTAIA domestic injury exception to apply and therefore adopt a proximate causation standard. As our sister circuits have explained, a proximate cause standard is consistent with prin- ciples of comity — “the respect sovereign nations afford each other by limiting the reach of their laws.” Empagran II, 417

IN RE DYNAMIC RANDOM ACCESS MEMORY 10655

F.3d at 1271 (internal quotations omitted). The Supreme Court counseled in Empagran I that the principles of comity require us to “ordinarily construe[ ] ambiguous statutes to avoid unreasonable interference with the sovereign authority of other nations.” 542 U.S. at 164; see In re Monosodium Glu- tamate, 477 F.3d at 538; Empagran II, 417 F.3d at 1271. To interpret the FTAIA broadly as requiring only “but for” cau- sation would risk the very sort of interference that we ordinar- ily seek to avoid. See Empagran I, 542 U.S. at 165 (“Why should American law supplant, for example, Canada’s or Great Britain’s or Japan’s own determination about how best to protect Canadian or British or Japanese customers from anticompetitive conduct engaged in significant part by Cana- dian or British or Japanese or other foreign companies?”).7

[8] Further, the Supreme Court said that “the FTAIA’s lan- guage and history suggest that Congress designed the FTAIA to clarify, perhaps to limit, but not to expand in any signifi- cant way, the Sherman Act’s scope as applied to foreign com- merce.” Empagran I, 542 U.S. at 169 (emphasis in original). If we were to adopt a “but for” standard, we would indeed be expanding the Sherman Act’s scope as applied to foreign commerce, notwithstanding Centerprise’s contention to the contrary.

None of the cases predating the FTAIA interpret the reach of the Sherman Act as expansively as Centerprise urges we should do here. The pre-FTAIA cases that Centerprise relies upon, Pfizer, Inc. v. Government of India, 434 U.S. 308 (1978), and Industria Siciliana Asfalti, Bitumi, S.p.A. v. Exxon Research & Engineering Co., No. 75 Civ. 5828-CSH, 1977 7 The case before us illustrates this point well. At oral argument, Center- prise acknowledged “[it] could” bring suit in the United Kingdom against the defendants for their anticompetitive conduct. We recognize that some of the defendants here are American corporations, but many are not, and Centerprise does not dispute that all of its purchases were made outside U.S. commerce, making real the risk of interference with a foreign nation’s ability to regulate its commercial affairs. 10656 IN RE DYNAMIC RANDOM ACCESS MEMORY WL 1353 (S.D.N.Y. Jan. 18, 1977), certainly do not reflect a “but for” standard. In Pfizer, although saying that Congress did not intend to make remedies under U.S. antitrust laws available only to U.S. consumers, the Supreme Court was speaking only to the issue of whether a foreign nation could be a “person” under antitrust laws. See Pfizer, 434 U.S. at 312, 314. It did not address the requisite causal relationship between the domestic effect and the foreign injury. If any- thing, the Court wrote in terms suggesting that the foreign sovereigns were directly injured by the defendants’ antitrust violations. See, e.g., id. at 314-15 (“To deny a foreign plaintiff injured by an antitrust violation the right to sue would . . . per- mit a price fixer or monopolist to escape full liability for his illegal actions and would deny compensation to certain of his victims, merely because he happens to deal with foreign cus- tomers.” (emphasis added)). In Industria Siciliana, the domes- tic effect at issue was plainly the direct cause of the plaintiff’s injury. There, a U.S. company was party to a reciprocal agree- ment that coerced the foreign consumer to forgo a more advantageous bid for design and engineering services from another U.S. company, a violation whose domestic effect directly and proximately caused the foreign plaintiff having to pay higher, super-competitive prices for the U.S. services. See Industria Siciliana, 1977 WL 1353, at *2, 11.

[9] Finally, the proximate cause standard is consistent with general antitrust principles, which typically require a direct causal link between the anticompetitive practice and plain- tiff’s damages. See Holmes v. Sec. Investor Prot. Corp., 503 U.S. 258, 267-68 (1992); Associated Gen. Contractors of Cal., Inc. v. Cal. State Council of Carpenters, 459 U.S. 519, 533-35 (1983); see also In re Monosodium Glutamate, 477 F.3d at 538-39. Accordingly, we hold that the “gives rise to” language of the domestic injury exception requires a direct or proximate causal relationship.

[10] Applying the proximate cause standard to Center- prise’s stated claim, we conclude that the domestic effect of

IN RE DYNAMIC RANDOM ACCESS MEMORY 10657

the defendants’ alleged price-fixing conspiracy did not give rise to Centerprise’s alleged foreign injury so as to satisfy the second prong of the FTAIA domestic injury exception. The defendants’ conspiracy may have fixed prices in the United States and abroad, and maintaining higher U.S. prices might have been necessary to sustain the higher prices globally, but Centerprise has not shown that the higher U.S. prices proxi- mately caused its foreign injury of having to pay higher prices abroad. Other actors or forces may have affected the foreign prices. In particular, that the conspiracy had effects in the United States and abroad does not show that the effect in the United States, rather than the overall price-fixing conspiracy itself, proximately caused the effect abroad. See In re Mono- sodium Glutamate, 477 F.3d at 539-40 (“The domestic effects of the price fixing scheme (increased U.S. prices) were not the direct cause of the appellants’ injuries. Rather, it was the for- eign effects of the price fixing scheme (increased prices abroad).” (emphasis added)).

[11] Notably, Centerprise is a foreign consumer that made its purchases entirely outside of the United States. It has recourse under its own country’s antitrust laws. See supra note 7. Centerprise’s indirectly linked foreign injury is not of the type that Congress intended to bring within the Sherman Act’s reach when it enacted the FTAIA, as evidenced by the domestic injury exception’s narrow phrasing. Centerprise’s causation theory directly parallels the “but for” or indirect causation theories rejected in Empagran II and In re Mono- sodium Glutamate.8 See Empagran II, 417 F.3d at 1270 8 Centerprise’s argument that its case is similar to Caribbean Broadcast- ing Systems, Ltd. v. Cable and Wireless PLC, 148 F.3d 1080 (D.C. Cir. 1998), and MM Global Services v. Dow Chemical Co., 329 F. Supp. 2d 337 (D. Conn. 2004), also fails. First, it is questionable whether Caribbean Broadcasting has any relevance, because it pre-dates Empagran I and does not interpret the provision of the FTAIA at issue here. See Caribbean Broad., 148 F.3d at 1087. Second, the case is factually distinguishable. There, the plaintiff alleged the defendants violated the Sherman Act in 10658 IN RE DYNAMIC RANDOM ACCESS MEMORY (rejecting as insufficient the plaintiffs’ theory that “[b]ecause the . . . product (vitamins) was fungible and globally mar- keted, [the defendants] were able to sustain super-competitive prices abroad only by maintaining super-competitive prices in the United States as well.”); In re Monosodium Glutamate, 477 F.3d at 539-40. Likewise, Centerprise’s statement that “[t]he United States prices were the source of, and substan- tially affected the worldwide DRAM prices” alleges no more than the “but for” causation that we hold does not suffice. In sum, Centerprise’s complaint suggests that super-competitive DRAM prices in the United States may have facilitated the defendants’ scheme to charge super-competitive prices abroad, but it does not sufficiently allege a theory that the higher U.S. prices proximately caused Centerprise’s foreign injury of having to pay higher prices outside the United States.

[12] Centerprise argues, however, that beyond alleging a theory similar to that rejected by our sister circuits, it is alleg- ing “a direct correlation between the U.S. price and the prices abroad” and “that the Defendants’ activities resulted in the U.S. prices directly setting the worldwide price.” At oral argu- ment, Centerprise was unable to explain how these additional allegations change its causation theory or how they make its

maintaining a broadcast monopoly in the eastern Caribbean by misrepre- senting to its advertising customers the breadth of the station’s broadcast- ing reach, an effect of which was that U.S. advertisers paid excessive prices for advertising. This effect in the United States directly caused the plaintiff to lose revenue because it could not sell advertising to the same U.S. businesses. Centerprise alleges no such foreclosure of the market by rivals. MM Global Services is also inapposite because the plaintiffs there pled direct participation in domestic commerce, unlike Centerprise whose complaint concerns only wholly foreign transactions. The plaintiffs argued that the defendants fixed minimum resale prices and other terms for sell- ing and reselling products in and from the United States, and compelled plaintiffs to agree to engage in a price maintenance conspiracy, thereby injuring the plaintiffs by precluding them from fully competing in the sales of their products. See MM Global Services, 329 F. Supp. 2d at 342.

IN RE DYNAMIC RANDOM ACCESS MEMORY 10659

claim distinguishable from those in Empagran II or In re Monosodium Glutamate. First, a direct correlation between prices does not establish a sufficient causal relationship. Simi- lar arguments were made and rejected in Empagran II and In re Monosodium Glutamate — that there was a single global price kept in equipoise by the maintenance of super- competitive prices in the U.S. market. See Empagran II, 417 F.3d at 1271, n.5; In re Monosodium Glutamate, 477 F.3d at 539-40. As to Centerprise’s assertion that the defendants’ activities resulted in the U.S. prices setting the worldwide price, Centerprise has taken liberties in characterizing the lan- guage of its complaint and, moreover, has not set forth a the- ory with any specificity of how this price-setting occurred or how it shows a direct causal relationship.9 See Kendall v. Visa U.S.A., Inc., 518 F.3d 1042, 1046-47 (9th Cir. 2008) (faulting antitrust complaint for failing to plead necessary evidentiary facts). We therefore hold that Centerprise has not sufficiently alleged that the domestic effect gave rise to its foreign injury so as to bring its claim within the FTAIA exception.

The district court properly concluded that Centerprise’s 9 Most notably, paragraph 75 of its complaint alleges: Memory purchases are a 24 hour global business, dependent in large part on United States events. For example, Plaintiff and many Class members start their days with communications to Defendants in Taiwan and Korea to understand what pricing is available for DRAM, and as the day goes on follow sales in the United States. Plaintiff and Class members were required to track the DRAM prices in dollars, which was the only available mea- sure due to Defendants’ sales and distribution practices, then work on dollar exchange rates in order to buy the DRAM at the best available price worldwide. The United States prices were the source of, and substantially affected the worldwide DRAM prices. The significance of these assertions, however, is not self-evident and Centerprise has not elaborated on how any of its asserted facts show that the higher U.S. DRAM prices proximately caused the excessive DRAM prices that Centerprise paid. 10660 IN RE DYNAMIC RANDOM ACCESS MEMORY failure to provide any comprehensible theory alleging a direct causal link between the domestic effects and the foreign injury warranted dismissal.

III. Leave to Amend

We review a district court’s denial of leave to amend a complaint for abuse of discretion. See McGlinchy v. Shell Chem. Co., 845 F.2d 802, 818 (9th Cir. 1988). A court prop- erly exercises its discretion in denying leave to amend if the proposed amendment would be futile. See Gabrielson v. Montgomery Ward & Co., 785 F.2d 762, 766 (9th Cir. 1986).

At oral argument before the district court, Centerprise pro- posed amending its complaint to include an allegation about the correlation between “U.S. prices during the conspiracy and those in Europe and Asia-Pacific.” Centerprise argued that this would show that the defendants’ “pricing of DRAM in the U.S. market, . . . with the knowledge that based on the way the market worked and the sales and distribution prac- tices, . . . was going to result in a direct correlation.” Center- prise argued this correlation was reflected in charts it submitted to the district court. The district court noted that an additional allegation about the correlation or interdependence of markets would not suffice to show that the effect in the United States actually “g[ave] rise to” the plaintiff’s claim; Centerprise could not escape a finding of no subject matter jurisdiction unless it came up with a different theory of recov- ery altogether.

[13] We agree that the additional allegation Centerprise proposed was similar to those it already made and alleged too indirect a link to support subject matter jurisdiction under the FTAIA. Centerprise’s assertion that the domestic and foreign prices were directly correlated, without more, did not warrant

IN RE DYNAMIC RANDOM ACCESS MEMORY 10661

granting leave to amend. Therefore, we hold the district court did not abuse its discretion in denying leave to amend as futile.10

The judgment of the district court is affirmed.

AFFIRMED.11

NOONAN, Circuit Judge, concurring:

There is such a thing as “cause in fact,” that is, an event that, not necessarily alone, brings about a given result. A “but for” cause is cause in fact. A “legal cause” is a cause in fact that is recognized in law as creating liability. A “proximate cause” is a legal cause. What turns cause in fact into a legal cause is a value judgment that the cause in fact creates an unacceptable risk of injury to a protected interest. What is an unacceptable risk and what is to be protected depend on the values of the society.

As Cardozo’s famous opinion in Palsgraf makes clear, that is how causation is often considered in the law of negligence. Antitrust law, apparently, still offers “proximate” as if it pro- vided something more than a label for a judgment that the conduct in question is foreseeably harmful to a social interest worthy of protection.

In the instant case, it would seem that reasonably prudent persons in the position of the defendants would see that their actions setting prices in the United States would negatively affect customers in the United States and elsewhere. But it has 10 In light of our decision on FTAIA subject matter jurisdiction and the district court’s refusal of leave to amend the complaint, we do not reach defendants’ contention that Centerprise lacks antitrust standing. 11 Appellees’ Request for Judicial Notice filed September 20, 2006 is denied as unnecessary. 10662 IN RE DYNAMIC RANDOM ACCESS MEMORY been the judgment of Congress and the Supreme Court that the economic interests of consumers outside the United States are normally not something that American law is intended to protect. Hence it is difficult to persuade a court that injury to foreign consumers has been “caused” by price-fixing in the United States. It’s so difficult that amendment of the com- plaint becomes futile and jurisdiction itself is found not to exist. We reach this vanishing point not from guidance in words like “proximate” or “direct” but from a strong sense that the protection of consumers in another country is nor- mally the business of that country. Location, not logic, keeps Centerprise’s claim out of court.