Opinion by Judge TALLMAN; Dissent by Judge ALDISERT.
TALLMAN, Circuit Judge:
We must decide whether the district court erred by determining that it lacked subject matter jurisdiction over this antitrust action. The United States alleged that an agreement between the defendants (collectively, “LSL”) and an Israeli company, Hazera Quality Seeds, Inc., violates the Sherman Act. Because the challenged agreement does not have a direct, substantial, and reasonably foreseeable effect on United States commerce, we affirm the district court’s dismissal.
I
This dispute grows out of a joint business venture — always a fertile ground for litigation — that sought to solve the dilemma of how to bring fresher, tastier tomatoes to Americans who live in the northern part of the nation and therefore suffer from a lack of fresh tomatoes in the winter months.
In the early 1980s, LSL Biotechnologies, Inc., an American corporation that develops and markets seeds, entered into a relationship with Hazera. LSL began working with Hazera in the hope of developing a genetically-altered tomato seed that would produce tomatoes with a longer shelf-life. LSL and Hazera wanted to create such a tomato because, until recently, tomatoes had a very short shelf-life if they were picked from the vine already ripened. This means that tomato growers can only sell their product in a limited geographic area. Because most of the American climate cannot produce tomatoes during the winter months, consumers are unable to access vine-ripened tomatoes for much of the year. Instead, most United States consumers are relegated to eating foreign tomatoes that are picked before they are ripe, so they will still be fresh after shipping. Tomatoes picked in this fashion have a poor flavor compared to vine-ripened tomatoes.
To solve this dilemma, LSL and Hazera sought to develop a tomato with enough shelf-life after reddening on the vine to travel from growing locations primarily in Mexico to the rest of the American market before spoiling. On January 1, 1983, LSL and Hazera signed a contract that regulated their relationship in this joint endeavor. The contract allocated to each party exclusive territories in which they could sell the seeds they developed together and seeds that each party developed on its own. The contract provided that LSL would have the exclusive rights to the North American market.
LSL and Hazera eventually bred a ripening-inhibitor (“RIN”) gene into tomato seeds to be grown in open fields. The RIN gene caused tomatoes to remain fresh longer after being picked. LSL obtained a patent for tomatoes and seeds containing the RIN gene; Hazera obtained no rights to the patent. The RIN gene tomatoes proved to be exceptionally successful when grown in Mexican climates, but failed to take in cooler American climates. As a result, Mexican growers now dominate the fresh winter-tomato market. To date, Hazera has not developed a long shelf-life tomato seed.
The relationship between LSL and Haz-era soon withered. Litigation ensued. In [675]*6751987, Hazera sued LSL in an Israeli court. This foreign litigation led to mediation in Israel that produced the renegotiation of and addendum to the contract. The addendum included a Restrictive Clause, which is the device the United States now claims violates the Sherman Act. The Restrictive Clause originally stated:
Subsequent to the termination of the Agreement hereunder, Hazera shall not engage, directly or indirectly, alone, with others and/or through third parties, in the development, production, marketing or other activities involving tomatoes having any long shelf life qualities. However, in the event that Hazera shall be requested by any third party to produce seeds of tomatoes having long shelf life qualities, Hazera may engage in such activities only if all of the following conditions are met: (A) the subject tomatoes do not have or involve long shelf life qualities which are included in LSL’s proprietary rights; (B) Hazera shall not engage in such production prior to the year 2000 or prior to the expiration of 5 years following the termination of the Agreement, whichever occurs later, and (C) Hazera has obtained LSL’s advanced written consent, which shall not be unreasonably withheld.... LSL shall determine whether or not the proposed cooperation may involve any of its proprietary rights and shall not unreasonably withhold its consents to such production.
LSL and Hazera continued working together after adopting the Restrictive Clause, despite frequent returns to the legal system. In 1992, the parties modified the contract a final time and requested that an Israeli arbitrator “incorporate their final contract modifications into a stipulated arbitration order.” The arbitration settlement affirmed the Restrictive Clause’s ban on Hazera selling long shelf-life tomato seeds in North America. But the Restrictive Clause was amended to allow Hazera to sell other seeds (e.g., tomato seeds for growing in greenhouses) to North American consumers, provided that Hazera disclose the details of such sales to LSL.
The contract between Hazera and LSL expired on January 1, 1996, and the Restrictive Clause became effective. On September 15, 2000, the United States filed its antitrust complaint. The government alleged that the Restrictive Clause is “so overbroad as to scope and unlimited as to time as to constitute a naked restraint of trade in violation of Section 1 of the Sherman Act.” The government also alleged that the Restrictive Clause is illegal because “it has harmed and will continue to harm American consumers by unreasonably reducing competition to develop better seeds for fresh-market, long shelf-life tomatoes for sale in the United States.”
The government alleged that “[b]ut for the[Restrictive Clause], Hazera would likely be a significant competitor of [LSL] in North America.” The Complaint stated that the defendants1 collectively held more than 70 percent of the market for “fresh market tomato seeds.” Nonetheless, LSL’s competitors control a significant percentage of the market for “fresh market tomato seeds”: Novartis and Monsanto together possess around twenty percent of the market, while several other companies together account for the remaining ten percent.
[676]*676The portion of the Complaint titled “Anticompetitive Effects” alleged that the exclusion of Hazera from the North American market eliminated “one of the few firms with the experience, track record and know-how likely to develop seeds that will allow United States and other North American farmers to grow better fresh-market tomatoes for United States consumers during winter months.” The government also alleged that the “Restrictive Clause may also allow defendants to charge more for their seeds (or more for a license to use seeds with the RIN gene) than they otherwise would.”
LSL filed a motion to dismiss the Complaint, arguing that the government failed to state a cause of action and that the district court lacked subject matter jurisdiction. In support of their positions regarding subject matter jurisdiction, the parties submitted declarations and other evidence. After hearing oral argument, the district court granted LSL’s motion.
The district court’s approach was to divide the Complaint into separate domestic and foreign components, because the area of restraint (North America) covered both domestic and foreign markets.
The district court first concluded that the Complaint failed to state a cause of action regarding conduct in the United States and dismissed that aspect of the action without prejudice under Federal Rule of Civil Procedure 12(b)(6). The court concluded that the Complaint’s market definition was so poor that the Complaint failed to establish anti-competitive effects. The United States chose not to amend its defective complaint.
The district court then held that it lacked jurisdiction over the claim that the restriction on the sale of seeds to Mexico violated the Sherman Act and dismissed that aspect of the Complaint with prejudice under Rule 12(b)(1). In order to seek immediate appellate review, the government, rather than replead the domestic conduct aspect of the Complaint, requested that the district court dismiss the entire action with prejudice. The request was granted and the government perfected its appeal to this court under 28 U.S.C. § 1291.
II
As an initial structural matter we must decide whether it was proper for the district court to divide the consideration of the Complaint into domestic and foreign components.
LSL moved to dismiss the Complaint, arguing primarily that the district court lacked subject matter jurisdiction. The motion very briefly urged that the whole Complaint—not just the domestic aspect— also failed to state a cause of action. LSL’s attack on the Complaint is not split into distinct “domestic” and “foreign” aspects. Likewise, the government’s opposition to LSL’s motion does not defend the Complaint in this fashion. The government concentrated the bulk of its response on the district court’s subject matter jurisdiction, while briefly answering LSL’s assertion that the entire Complaint failed to state a cause of action.
The district court evaluated different parts of the Complaint under different standards. On the one hand, the court considered whether the allegations in the Complaint concerning the restriction on selling seeds to the United States stated a cause of action under Rule 12(b)(6). The district court was satisfied that it had subject matter jurisdiction over this component of the Complaint. On the other hand, the district court analyzed whether the court had subject matter jurisdiction over the portion of the Complaint alleging a restraint on the sale of seeds to Mexico [677]*677under Rule 12(b)(1) and the Foreign Trade Antitrust Improvements Act of 1982, 15 U.S.C. § 6a (“FTAIA”).
Of course, the parties’ arguments did not prevent the district court from separately considering the Complaint’s domestic and foreign allegations. Nonetheless, we think this two-pronged approach was not the most appropriate way to analyze the motion to dismiss. The threshold question in this case is whether the district court had subject matter jurisdiction. That inquiry pervades the entire Complaint; the government alleged only one cause of action, which lumped together the Restrictive Clause’s ban on distributing certain modified tomato seeds to Mexico and the resulting fruit in the United States.
Where, as here, a Complaint alleges a restraint of trade on a foreign corporation, that restraint was executed in a foreign nation as the result of litigation in that foreign nation, and the defendants file a Rule 12(b)(1) motion to dismiss the entire Complaint for lack of subject matter jurisdiction, a court should determine its subject matter jurisdiction to entertain the entire Complaint. Accordingly, we consider whether the district court had subject matter jurisdiction over the entire Complaint, including the “domestic” allegations.
Ill
We review de novo the district court’s dismissal for lack of subject matter jurisdiction. La Reunion Francaise SA v. Barnes, 247 F.3d 1022, 1024 (9th Cir.2001). Factual findings made in support of the dismissal are reviewed for clear error. Id.
IV
The Sherman Act prohibits “[ejvery 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. Federal courts have struggled for decades to determine when United States courts have jurisdiction over allegations of foreign restraints of trade. See Areeda & Hovenkamp, ANTITRUST LAW, ¶ 272 (2d ed.2000); see also Den Norske Stats Oljeselskap As v. HeereMac v.o.f., et al., 241 F.3d 420, 423-24 (5th Cir.2001) (“The history of this body of case law is confusing and unsettled.”).
Prior to the passage of the FTAIA, courts applied varying tests to determine when foreign conduct fell within the purview of the Sherman Act. The most widely used standard was the “effects test,” which was developed by Judge Learned Hand in United States v. Aluminum Co. of Am., 148 F.2d 416, 444 (2d Cir.1945) (‘Alcoa’’).2 The Alcoa court considered whether Congress intended the Sherman Act to impose liability for conduct outside of the United States and whether the Constitution allowed Congress to do so. Judge Hand rejected the idea that Congress meant “to punish all whom[United States] courts can catch, for conduct which has no consequences within the United States.” Id. at 443. Instead, the court held that the Sherman Act was meant to reach foreign conduct only if it was intended to and did affect United States commerce. Id. In Hartford Fire Ins. Co. v. California, the Supreme Court summarized the effects test, stating that “it is well established by now that the Sherman Act applies to foreign conduct that was meant to produce and did in fact produce some substantial [678]*678effect in the United States.” 509 U.S. 764, 796, 113 S.Ct. 2891, 125 L.Ed.2d 612 (1993) (citing Alcoa).
Application of the effects test, however, has proved difficult and the precise extraterritorial reach of the Sherman Act remains less than crystal clear. In an effort to address this uncertainty, Congress enacted the FTAIA in 1982. The FTAIA “was intended to exempt from the Sherman Act export transactions that did not injure the United States economy.” Hartford Fire, 509 U.S. at 796-97 n. 23, 113 S.Ct. 2891. According to the House Report, another significant purpose of the FTAIA was to fix the problem that arose because “courts differ in their expression of the proper test for determining whether United States antitrust jurisdiction over international transactions exists.” H.R.Rep. No. 97-686 (1982), reprinted in 1982 U.S.C.C.A.N. 2487 (“House Report”) at 2487. The FTAIA tackled this issue by “clarifying the Sherman Act ... to make explicit [its] application only to conduct having a ‘direct, substantial and reasonably foreseeable effect’ on domestic commerce.” Id. Specifically, the FTAIA states:
Sections 1 to 7 of this title [including 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 foreseeable 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 he provisions of sections 1 to 7 of this title, other than this section.
15 U.S.C. § 6a.
Federal courts did not shower the FTAIA with attention for the first decade after its enactment. But in the last ten years, and in particular the last five years, the case reporters have steadily filled with decisions interpreting this previously obscure statute. As a threshold matter, many courts have debated whether the FTAIA established a new jurisdictional standard or merely codified the standard applied in Alcoa and its progeny.
Several courts have raised this question without answering it. The Supreme Court did as much in Hartford Fire. The Hartford Fire Court considered the jurisdictional status of an alleged conspiracy among London-based re-insurers to manipulate the American insurance market by not offering certain types of re-insurance. The Court introduced its jurisdictional analysis by noting that the effects test is well established. Hartford Fire, 509 U.S. at 796, 113 S.Ct. 2891. In a footnote, the Court stated that it is not clear “whether the [FTAIA’s] ‘direct, substantial, and reasonably foreseeable effect’ standard amends existing law or merely codifies it. We need not address these questions here.” Id. at 796 n. 23, 113 S.Ct. 2891.3
[679]*679It is manifest that our role is to apply the laws that Congress passes and the executive branch enforces unless those laws violate the Constitution. There is no suggestion that the FTAIA is unconstitutional. Thus, we must adhere to the FTAIA in determining whether a district court has subject matter jurisdiction over an alleged foreign restraint of trade. The government contends that the FTAIA merely codified the existing common law regarding when the Sherman Act applies to foreign conduct and that we should continue to employ the Alcoa effects test. We reject this contention.
Our task when interpreting legislation is to give meaning to the words used by Congress; we strive to avoid constructions that render words meaningless. See United States v. Fiorillo, 186 F.3d 1136, 1153 (9th Cir.1999). The FTAIA states that the Sherman Act shall not apply to foreign conduct unless it has a “direct, substantial, and reasonably foreseeable effect” on domestic commerce. 15 U.S.C. § 6a(l). The Supreme Court reads the Alcoa test as conferring jurisdiction so long as the conduct creates “some substantial effect in the United States.” Hartford Fire, 509 U.S. at 796, 113 S.Ct. 2891. Unlike the FTAIA, the Alcoa test does not require the effect to be “direct.” Adopting the government’s argument and applying the Alcoa test would render meaningless the word “direct” in the FTAIA.4 We are not willing to rewrite a statute under the pretense of interpreting it.
Moreover, applying Alcoa instead of the FTAIA would contravene the FTAIA’s purpose. The FTAIA created its jurisdictional test because the “enactment of a single, objective test — the ‘direct, substantial, and reasonably foreseeable effect’ test — will serve as a simple and straightforward clarification of existing American law.” House Report at 2487-88. The House Report goes on to state: “The specific purpose of the Sherman Act modification is: to more clearly establish when antitrust liability attaches to international business activities.” Id. at 2492.
It would be a serious departure from the goal of achieving clarity for us to conclude that Congress meant only “some substantial effect,” Hartford Fire, 509 U.S. at 796, 113 S.Ct. 2891, when it said “direct, substantial, and reasonably foreseeable effect.” Clarity is not achieved by employing three modifiers (“direct,” “substantial,” and “reasonably foreseeable”) as the standard for the required effect of the challenged conduct and then telling businesses that only one modifier (“substantial”) is relevant to Sherman Act liability.
Our precedent supports the conclusion that the FTAIA provides the guiding standard for jurisdiction over foreign restraints of trade. Most notably, in McGlinchy v. Shell Chem. Co., 845 F.2d 802 (9th Cir.1988), we considered an antitrust claim that the defendants’ refusal to deal in pipe-manufacturing products “in various foreign markets” violated the Sherman Act. Id. at 813. The district court dismissed the claim because it failed to satisfy the FTAIA’s test for subject matter jurisdiction. We determined without difficulty that “we are bound to apply [the FTAIA],” id. at 813 n. 8, and affirmed the district court, concluding that “appellants have failed to allege that the defendants’ conduct has a ‘direct, substantial, and reasonably foreseeable effect’ on domestic commerce or import trade.” Id. at [680]*680815.5
As in McGlinchy, we conclude that the FTAIA controls in this case. Therefore, we must affirm the district court’s dismissal for lack of subject matter jurisdiction unless we determine that the Restrictive Clause operates to have a direct, substantial, and reasonably foreseeable effect on domestic commerce.6
V
Before discussing whether there is a direct effect here, we must consider what Congress meant by “direct.” A dictionary published contemporaneously with the enactment of the FTAIA defined “direct” as “proceeding from one point to another in time or space without deviation or interruption.” WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 640 (1982).
Further, our efforts at understanding the meaning of “direct” are aided by the Supreme Court’s interpretation of a nearly identical term in the Foreign Sovereign Immunities Act (“FSIA”). The FSIA states that immunity does not extend to commercial conduct “outside the territory of the United States ... that [ ] causes a direct effect in the United States.” 28 U.S.C. § 1605(a)(2). After the lower federal courts struggled for years to define “direct effect,” the Supreme Court unanimously declared that an effect is “direct” if it follows as an immediate consequence of the defendant’s activity. Republic of Argentina v. Weltover, Inc., 504 U.S. 607, 618, 112 S.Ct. 2160, 119 L.Ed.2d 394 (1992). Settling on this definition, the Court “reject[ed] the suggestion that [‘direct’] contains any unexpressed requirement of ‘substantiality’ or ‘foreseeability.’ ” Id.
Having defined “direct,” we next consider what effects the government asserts. As the district court recognized, the Complaint alleges that the Restrictive [681]*681Clause causes two effects: (1) the agreement makes less likely possible innovations from Hazera in the creation of heartier tomato seeds “that will allow consumers to enjoy higher quality, better tasting winter tomatoes and that will allow United States farmers to grow long shelf-life tomatoes,” and (2) the Restrictive Clause “may also allow defendants to profitably charge more for their seeds (or more for a license to use seeds with the RIN gene) than they otherwise could.”
Neither of these effects is “direct.” The delay of possible “innovations” does not have a direct effect on American commerce. Even if Hazera were free to distribute new types of long shelf-life seeds in North America, there is no indication that Hazera has yet figured out a different way to produce such a seed without violating LSL’s intellectual property rights to the RIN gene. Moreover, there is no indication that Hazera is in a stronger (or as strong a) position to develop such a new seed as Monsanto and Novartis, which the Complaint alleges already account for 20 percent of the tomato seed market. Thus, any innovation that Hazera would bring to American consumers is speculative at best and doubtful at worst. An effect cannot be “direct” where it depends on such uncertain intervening developments. In this case, Hazera’s delivery of long shelf-life seeds to North American growers depends on Hazera first creating such seeds, a development that is certainly not guaranteed.7
We can imagine a situation where the exclusion of a potential foreign competitor would satisfy the “direct” requirement. One such scenario might be where the foreign competitor already has the good in hand. It might also be possible for a “direct” effect to exist where the potential foreign competitor does not yet have the product in hand. A potential foreign competitor might be able to demonstrate that its exclusion already has an effect on the American market. For example, the competitor might be able to demonstrate that its exclusion is causing existing market players to invest less in the research and development of new products. Although it might be possible in such situations for a [682]*682“direct” effect to exist, the United States has not presented us with sufficient evidence to conclude that the district court clearly erred in ruling on the existing pleadings that Hazera’s exclusion does not yet have a direct effect on domestic commerce.
The district court also held that the effect of the Restrictive Clause on prices paid by American consumers is not “direct.” This ruling was not clearly erroneous. The government has presented no evidence that LSL has or will artificially inflate the prices it charges to Mexican farmers for LSL’s long shelf-life seeds.
Perhaps more importantly, the government’s argument about the directness of the effect of seed prices on tomato prices is undermined by the United States’ recent agreement with Mexican farmers to set a floor on the price of tomatoes shipped from Mexico to the United States. SUSPENSION OF ANTIDUMPING INVESTIGATION: FRESH TOMATOES FROM MEXICO, 67 Fed.Reg. 77044 (Dec. 16, 2002). The agreement covers all “fresh or chilled tomatoes (fresh tomatoes) which have Mexico as their origin.” Id. at 77046. The agreement fixes minimum prices that Mexican tomato growers must charge for their product. Id. at 77045-50. The agreement was necessary to “eliminate completely the injurious effect of exports to the United States of the subject merchandise and prevent the suppression or undercutting of price levels of domestic fresh tomatoes by imports of that merchandise from Mexico.” Id. at 77045. The government initiated the investigation that led to the agreement because it was determined that Mexican tomato growers were selling their product “in the United States at less than fair value.” Id. This agreement, and the concerns that gave rise to it, belie the United States’ argument that LSL could raise the prices ultimately paid by American tomato consumers.
The government cites two cases in support of its argument that the Restrictive Clause has a “direct” effect on American commerce. First, the government points to Hartford Fire. However, the defendants in Hartford Fire “apparently eoncede[d]” that the district court had jurisdiction; rather than challenge the existence of jurisdiction, the defendants argued that the district court should decline to exercise jurisdiction because of comity concerns. 509 U.S. at 795, 113 S.Ct. 2891. Also, the Hartford Fire defendants’ foreign conduct — not offering re-insurance — had a demonstrated impact on the American insurance market: certain types of primary insurance were made unavailable because primary insurers could not obtain necessary re-insurance. Here, the product “loss” suggested by the government is entirely speculative: but for the Restrictive Clause, Hazera might someday create a long shelf-life tomato seed suitable for growing by North American farmers that does not violate LSL’s existing patents. While the Restrictive Clause in this case removes nothing from American consumers except the vague possibility that Haz-era might create a new type of seed before Novartis or Monsanto or one of the other smaller competitors, conduct by the defendants in Hartford Fire actually deprived American consumers of a wider array of an existing product, primary insurance.
The government also relies on the Fifth Circuit’s decision in Den Norske. Like Hartford Fire, Den Norske differs in at least one critical way from this case. In Den Norske, the plaintiff alleged that the defendants caused Americans to pay $165 million in higher oil prices because the defendants’ conspiracy raised the prices paid by off-shore oil producers for heavy-lift services in the Gulf of Mexico and the [683]*683producers passed on the higher prices to American oil consumers. 241 F.3d at 426 & n. 21. By contrast, here the government cannot demonstrate an existing effect on American tomato consumers. At most, the government can demonstrate that the Restrictive Clause removes the possibility of future innovation from Haz-era and “may also allow defendants to profitably charge more for their seeds.”
In assailing the district court’s jurisdictional ruling, the dissent invokes the general notice pleading standard. See, e.g., Dissent at 692-93 n. 5 (“Here, the government sufficiently averred that the adverse effect on the United States’ domestic commerce of tomatoes has been an immediate consequence of the Restrictive Clause governing tomato seeds .... ”); id. at 694 (“First, the Complaint squarely alleged causation in fact: but for the restraint, United States consumers would have the important potential of better winter tomatoes grown from Hazera seeds.”). However, while federal complaints are generally construed liberally, in this case the district court correctly scrutinized the government’s Complaint more closely in order to make the necessary threshold determination of whether it had subject matter jurisdiction. This scrutiny involved taking preliminary evidence, weighing that evidence, and deciding as a matter of law whether the facts alleged supported jurisdiction. Here, the district court properly found that the operation of the Clause does not have the required direct effect on U.S. commerce. We agree and hold that as a matter of law the FTAIA does not confer jurisdiction.8
VI
The FTAIA provides the standard for establishing when subject matter jurisdiction exists over a foreign restraint of trade. This standard was not met here because the government cannot demonstrate that the district court clearly erred by determining that the alleged effects of the Restrictive Clause are not direct.
Because we conclude that the district court lacked subject matter jurisdiction over the entire Complaint, we do not consider the district court’s Rule 12(b)(6) dismissal of the domestic aspect of the Complaint.
AFFIRMED.