OPINION

After this court in an unpublished disposition affirmed summary judgment in his favor as a defendant in a federal court action alleging violations of the Securities Exchange Act of 1934, 15 U.S.C. §§ 78j(b), 78t(a), and RICO, 18 U.S.C. §§ 1961-1968, Carl Berg sued the plaintiff in that action, Hayden Lea-son, and Leasoris lawyers, Heller, Ehrman, White & McAuliffe, ^[1] for malicious prosecution in state court. Leason removed on the ground that the malicious prosecution claim was based on alleged violations of federal law and therefore “arises under” the laws of the United States for purposes of federal question jurisdiction. ^[2] 28 U.S.C. §§ 1331, 1441(b). The district court declined to remand in a published opinion. Berg v. Leason, 793 F.Supp. 930 (N.D.Cal.1992). We hold that the federal element is insufficiently substantial to confer “arising under” jurisdiction because the malicious prosecution court need only decide whether the underlying claim was “legally tenable,” the cause of action is created by state law, and state law controls the standard by which the strength of the federal claim in the underlying action is measured. We therefore reverse. ^[3]

I

The rule is well settled that a state claim “arises under” federal law “if the complaint, properly pleaded, presents a substantial dispute over the effect of federal law, and the result turns on the federal question.” Guinasso v. Pacific First Fed. Sav. & Loan Ass’n, 656 F.2d 1364, 1365-66 (9th Cir.1981), cert. denied, 455 U.S. 1020, 102 S.Ct. 1716, 72 L.Ed.2d 138 (1982). The “vast majority of cases brought under the general federal-question jurisdiction of the federal courts are those in which federal law creates the cause of aetion[,]” Merrell Dow Pharmaceuticals Inc. v. Thompson, 478 U.S. 804, 808, 106 S.Ct. 3229, 3232, 92 L.Ed.2d 650 (1986), but a case may also arise under federal law “ ‘where the vindication of a right under state law necessarily turn[s] on some construction of federal law.’ ” Id. (quoting Franchise Tax Bd. v. Construction Laborers Vacation Trust, 463 U.S. 1, 9, 103 S.Ct. 2841, 2846, 77 L.Ed.2d 420 (1983)).

We have not had occasion to apply these rules in the context of a suit for malicious prosecution where the underlying action arose under federal law and was tried in federal court. ^[4] However, we considered an[*424] analogous situation in Hunter v. United Van Lines, 746 F.2d 635 (9th Cir.1984), cert. denied, 474 U.S. 863, 106 S.Ct. 180, 88 L.Ed.2d 150 (1985). There, plaintiffs brought a number of claims in state court against a common carrier which were removed to federal court, including one for bad faith handling of damage claims. That claim depended on a contract claim governed by federal law (the Carmack Amendment, 49 U.S.C. § 11707), and so federal law was “an ingredient” in plaintiffs state law claim for tortious bad faith. Id. at 645. As in this case, the defendants argued that federal law was an element of the plaintiffs state cause of action and for that reason the action was one “aris[ing] under” federal law. In deciding that the state claim was not transformed into a federal claim just because federal law plays a preliminary, threshold role, we drew on the Supreme Court’s seminal opinion in Gully v. First National Bank, 299 U.S. 109, 57 S.Ct. 96, 81 L.Ed. 70 (1936), to explain:

Justice Cardozo instructed us to determine whether the federal element in the claim was “basic” as opposed to “collateral,” and “necessary” as opposed to “merely possible.” Gully, 299 U.S. at 118, 57 S.Ct. at 100. Similarly, courts have looked to whether the federal element in the claim was “pivotal,” or “substantial,” as opposed to merely “incidental,” or whether it was “direct and essential” as opposed to “attenuated,” or “paramount” as opposed to “collateral.” Thus, the resolution of the federal question must play a significant role in the proceedings.

Hunter, 746 F.2d at 646 (citations and alteration omitted). Because California’s duty of good faith was triggered by a “colorable claim, as opposed to one ultimately found valid,” we held that the federal element was insufficiently substantial under any standard. ^[5] Id.

Guided by Hunter, and mindful of “the need for careful judgments about the exercise of federal judicial power in an area of uncertain jurisdiction,” Merrell Dow, 478 U.S. at 814, 106 S.Ct. at 3235, we turn to Berg’s claim in this case.

II

To prove malicious prosecution in California, a plaintiff “must demonstrate that the prior action (1) was commenced by or at the direction of the defendant and was pursued to a legal termination in his, plaintiff’s, favor; (2) was brought without probable cause; and (3) was initiated with malice.” Sheldon Appel Co. v. Albert & Oliker, 47 Cal.3d 863, 871, 254 Cal.Rptr. 336, 765 P.2d 498 (1989) (quotation marks and alterations omitted).

Berg’s amended complaint alleges a single count for malicious prosecution based on the final resolution in Berg’s favor of Leason’s action against Berg in the United States District Court for the Northern District of California. It alleges that Leason and his counsel “acted without probable cause in initiating and maintaining the [underlying action] in that the claims were not tenable under the relevant law and facts at the time the action was brought.” The underlying action had to do with Leason’s investing in Actrix Computer Co., of which Berg was a director and major shareholder, and sought damages on eight different causes of action, including federal securities fraud and violation of RICO.

Leason rests his jurisdictional argument on the fact that the relevant law in the underlying action was federal securities law and federal RICO law. He argues that because the court will have to analyze the federal securities and RICO claims and whether probable cause supported them, pivotal and substantial questions of federal law are necessarily raised. We disagree.

Berg’s burden is not to show that RICO, or the securities laws, were or were[*425] not violated; that, of course, was favorably determined in the underlying action. Rather, “the probable cause element calls on the trial court to make an objective determination of the ‘reasonableness’ of the defendant’s conduct, i.e., to determine whether, on the basis of the facts known to the defendant, the institution of the prior action was legally tenable.” Sheldon Appel, 47 Cal.3d at 878, 254 Cal.Rptr. 336, 765 P.2d 498. “Legally tenable” is a low level of generality. It follows from Hunter, where we said that no substantial federal question is raised when the state-law claim is triggered by a “color-able” federal claim, Hunter, 746 F.2d at 646, that the federal element in an action for malicious prosecution is not substantial when the probable cause threshold is crossed by a “legally tenable” claim.

Normally, as Justice Holmes put it, “[a] suit arises under the law that creates the cause of action.” American Well Works Co. v. Layne & Bowler Co., 241 U.S. 257, 260, 36 S.Ct. 585, 586, 60 L.Ed. 987 (1916); see also Merrell Dow, 478 U.S. at 808, 106 S.Ct. at 3232. The cause of action for malicious prosecution is of course a creature of state law. State law controls the other elements which must be proved, as well as what suffices for probable cause. Thus, in California, the probable cause inquiry has two components, one factual and one legal. Before reaching the question of whether the underlying action was legally tenable, the factfinder must determine what the defendant knew or believed about the facts. Sheldon Appel, 47 Cal.3d at 881, 254 Cal.Rptr. 336, 765 P.2d 498; see also Leonardini v. Shell Oil Co., 216 Cal.App.3d 547, 568, 264 Cal.Rptr. 883 (1989) (probable cause analysis differs depending on whether alleged deficiency lies in relying on facts litigant has no reasonable cause to believe are true or seeking recovery on legal theory which is untenable on the facts that were known), cert. denied, 498 U.S. 919, 111 S.Ct. 293, 112 L.Ed.2d 247 (1990). Neither the factual part of the probable cause element, nor the separate question of malice, turns at all on federal law.

Finally, the court looks at the merits of a claim for malicious prosecution through the prism of state law. For example, malicious prosecution actions are disfavored in California, Sheldon Appel, 47 Cal.3d at 872, 254 Cal.Rptr. 336, 765 P.2d 498; Leonardini, 216 Cal.App.3d at 566, 264 Cal.Rptr. 883, and underlying complaints are construed liberally in favor of the plaintiff in that action instead of restrictively in favor of the malicious prosecution plaintiff. Leonardini, 216 Cal.App.3d at 571, 264 Cal.Rptr. 883. More significantly, how much merit the underlying action must have is a state-law matter. As we recognized in Hunter, federal law cannot be controlling when the degree of substance in the federal claim necessary to trigger the state-law cause of action is a question of state law. Hunter, 746 F.2d at 646. Likewise here, federal law is not dispositive because the degree of strength required to put the underlying federal claim over the probable cause threshold is determined by state law.

We therefore conclude that the task of deciding whether an underlying federal claim was legally tenable does not raise a substantial question of federal law.

Ill

Leason argues that we should conclude differently out of concern for the right to seek redress for violations of federal law in federal court. He first suggests that malicious prosecution actions could be used improperly to inhibit plaintiffs from bringing legitimate actions under federal law. We do not see how: Two federal courts have already determined that Leason’s federal claims lacked merit. The federal interest, which lies in providing a forum for the protection of federally-created rights, has already been served and will not be disserved by a state court determining whether the federal complaint met the test of being “legally tenable.”

Leason next points out that Congress gave the federal courts exclusive jurisdiction over the 1934 Act in order to secure uniformity in its interpretation. 15 U.S.C. § 78aa. From this he posits that federal judges know more about the securities laws than do state judges, and that malicious prosecution suits in state courts might defeat federal policy. While this argument has[*426] greater force in the case of the underlying securities claim than the RICO claim (where state courts have concurrent jurisdiction), we are not persuaded. State courts resolve matters of federal law in similar circumstances with no difficulty; neither an affirmative defense based on federal law, Merrell Dow, 478 U.S. at 808, 106 S.Ct. at 3232; Weiner v. Shearson, Hammill & Co., 521 F.2d 817, 822 (9th Cir.1975), nor one based on federal preemption, Caterpillar Inc. v. Williams, 482 U.S. 386, 393, 107 S.Ct. 2425, 2430, 96 L.Ed.2d 318 (1987), renders an action brought in state court removable. In these circumstances state courts directly decide issues of federal law, yet their capacity to do so is inherently a part of the well-pleaded complaint rule. See Franchise Tax Bd., 463 U.S. at 9-12, 103 S.Ct. at 2846-47. We therefore see no reason why concern for consistency of federal law — which animates the doctrine of preemption as well as the doctrine of exclusive jurisdiction — should convert a state cause of action with only a tangential federal element into a substantial federal question.

Finally, Leason argues that federal jurisdiction is essential for evaluating claims brought under federal law where the scope of the federal law is still evolving. This argument also fails because by definition, the scope of the federal law will have been tested and determined in the federal court action. A state court can decide as competently as can a federal court whether the threshold of a “legally tenable” claim existed.

Accordingly, we hold that Berg’s malicious prosecution action does not “arise under” federal law simply because one element requires proof that the underlying federal action was legally untenable. As there is no original federal jurisdiction under 28 U.S.C. § 1331, removal jurisdiction was improperly exercised and the action must be remanded to state court.

REVERSED AND REMANDED.