dissenting:
I respectfully dissent.
We must decide whether removal is proper under the “mass action” provision of the Class Action Fairness Act of 2005 (“CAFA”), Pub.L. No. 109-2, 119 Stat. 4 (2005), when plaintiffs move for coordination pursuant to California Code of Civil Procedure section 404 and justify their request in part by asserting a need to *925avoid inconsistent judgments.1 CAFA extends federal removal jurisdiction for mass actions when “monetary relief claims of 100 or more persons are proposed to be tried jointly on the ground that the plaintiffs’ claims involve common questions of law or fact.” 28 U.S.C. § 1332(d)(ll). I would hold that these requirements are met, and would reverse the district court’s remand order.
I
The issue before us is whether Plaintiffs’ petition to coordinate actions under California Code of Civil Procedure section 404 constitutes a proposal for these actions in California state court to be tried jointly, making the actions a “mass action” subject to federal jurisdiction under CAFA. I agree with the majority that federal courts are courts of limited jurisdiction, and the general rule is that removal statutes are strictly construed against removal.2 Luther v. Countrywide Home Loans Servicing LP, 533 F.3d 1031, 1034 (9th Cir.2008). As such, I turn to the language and purpose of CAFA. The statutory issue for us is whether the petition that was filed in this case seeking coordination of the California propoxyphene actions was a proposal in substance for those actions to be tried jointly. I regret that the majority here misinterprets CAFA and does so in a way that creates a circuit split, for practical purposes, with the Seventh Circuit’s decision in Abbott.
Congress enacted CAFA in 2005 to “curb perceived abuses of the class action device which, in the view of CAFA’s proponents, had often been used to litigate mul-ti-state or even national class actions in state courts.” Tanoh v. Dow Chemical Co., 561 F.3d 945, 952 (9th Cir.2009) (citation omitted). CAFA further extends federal jurisdiction over “mass action” cases when several requirements are met, although only the “proposed to be tried jointly” requirement is at issue here. See 28 U.S.C. § 1332(d)(2), (6), (11)(A).
Proposals for joint trials may be made implicitly, and a “joint trial” may “take different forms as long as the plaintiffs’ claims are being determined jointly.” Abbott, 698 F.3d at 573; see Bullard v. Burlington N. Santa Fe Ry. Co., 535 F.3d 759, 762 (7th Cir.2008). For example, an “exemplary” or “bellwether” trial may only feature a small group of plaintiffs, but it is still a joint trial when the claims or issues of a larger group are precluded or otherwise decided by the results. See Koral v. Boeing, Co., 628 F.3d 945, 947 (7th Cir.2011). We should be looking at the reality of joint trial proposal, not at how a party may characterize its own actions.
What is critical is that this appeal concerns a set of actions filed in state court followed by a petition by Plaintiffs to coordinate, in part to avoid inconsistent judgments. And so it is on that aspect of this case, distinguishing it from Tanoh, that we should be focused.3
*926My disagreement with the majority is over the import of the coordination motion and the reasons given for it. The majority focuses on the part of the petition mentioning pretrial discovery and chooses to downplay that part of the petition urging that there be no inconsistent judgments. In doing this, the majority disregards that the proviso in CAFA makes clear only that matters consolidated exclusively for pretrial purposes are not properly removed to federal court. The majority does not try even to argue, nor could it do so correctly here, that the petition for coordination is limited to pretrial matters. Instead, it argues that the petition “stopped far short of proposing a joint trial.” But there is no applicable judicial precedent supporting the majority’s proposition that the focus of a coordination petition mentioning pretrial matters in large part may override the reality of a plaintiffs proposal to try claims jointly when the petition seeks relief that would require joint trial. The majority apparently would require an explicit request for a joint trial, whereas I conclude that the substance of what was done is controlling. Recourse to the general principle that doubts on removal should be resolved by favoring the plaintiffs’ forum choice simply does not answer that this case fits CAFA removal like a glove under a reasonable assessment of what is a proposal for joint trial.
Our Ninth Circuit precedent in Tanoh suggests that plaintiffs are the “masters of their complaint,” and do not propose a joint trial simply by structuring their complaints so as to avoid the one hundred-plaintiff threshold. 561 F.3d at 953, 956; see Anderson v. Bayer; 610 F.3d 390, 393 (7th Cir.2010); Scimone v. Carnival Corp., 720 F.3d 876 (11th Cir.2013). That is not surprising and is analogous to the fact that individuals and corporations can structure transactions so as to avoid statutory prohibitions or terms.
But the United States Supreme Court has recently pointed out that there are limits to how far plaintiffs may go in structuring their complaints to avoid federal jurisdiction. Thus in Standard Fire v. Knowles, the Supreme Court rejected the ability of a proposed class action plaintiff to stipulate that damages would not exceed five million dollars. 568 U.S. -, 133 S.Ct. 1345, 1350, 185 L.Ed.2d 439 (2013) (“[T]he stipulation at issue here can tie Knowles’ hands, but it does not resolve the amount-in-controversy question in light of his inability to bind the rest of the class.”). In that case, the plaintiff unsuccessfully attempted to stipulate an amount-in-controversy below five million dollars before his proposed class had been certified. Id. at 1347. Standard Fire arose in the context of a challenge to plaintiffs’ counsel’s attempt to limit damages before class certification, and the Court recognized that plaintiffs’ counsel could not execute a damages stipulation binding class claimants not yet joined. So Standard Fire is in my view not necessarily controlling on the issue before us as to whether there has been a proposal for joint trial. Because in Standard Fire the Supreme Court appeared to reiterate that plaintiffs are the “masters of their complaint,” id. at 1350, if Plaintiffs merely had structured separate actions with less than one hundred claimants, and did not seek to coordinate them, I must currently think that the Supreme Court would hold, as we did in Tanoh, that no *927mass action was presented. If plaintiffs are masters of their complaints and can plead in a way to avoid federal jurisdiction, they remain free to “game” the system to some degree, including by joining less than one hundred plaintiffs in many suits in state court, so long as those cases are separate. Nonetheless, we have in this case a request to California courts to coordinate the actions and reasons given for coordination, including to avoid inconsistent judgments. That leads me to recognize that the issue here, stated more precisely, is whether when plaintiffs seek to coordinate under California law many state actions, and urge the state court that coordination is necessary to avoid inconsistent judgments, that is a proposal for joint trial within the meaning of CAFA.
Plaintiffs argue, and the majority agrees, that their petition for coordination was not a proposal to try the cases jointly. I must respectfully disagree. California Code of Civil Procedure section 404 allows the coordination of “all of the actions for all purposes,” and presents a factor-based test to determine whether coordination is appropriate. Plaintiffs asked for coordination under section 404, and submitted a memorandum in support of the petition for coordination. Reasons Plaintiffs listed as supportive of their petition, including the danger of inconsistent judgments and conflicting determinations of liability, in my view could only be addressed through some form of joint trial. When Plaintiffs asked the California Judicial Council to coordinate their cases for reasons that only a joint trial could address, they implicitly proposed a joint trial, bringing their cases within CAFA’s mass action provision.4 That is how I see it and that is what impels my dissent.
Plaintiffs further contend that we should interpret the phrase “joint trial” to mean “a joint trial where more than one party (and for purposes of CAFA 100 or more parties) simultaneously present their claims to a trier of fact.” I would reject this interpretation because it violates the canon against reading a statutory provision in such a way as to render another provision superfluous. See Bilski v. Kappos, — U.S. -, 130 S.Ct. 3218, 3228, 177 L.Ed.2d 792 (2010) (citation omitted). If our court were to adopt Plaintiffs’ interpretation of “joint trial,” the mass action statutory exception for “claims [that] have been consolidated or coordinated solely for pretrial proceedings” would be meaningless because a proposal for anything short of a single massive trial for all claimants would already fail the mass action requirement. 28 U.S.C. § 1332(d)(ll)(B)(ii).5 I would reject Plaintiffs’ narrow interpretation of “joint trial” to give meaning to the exception above.
Although Plaintiffs argue that the Seventh Circuit decision in Abbott is inapplica*928ble here, and the majority accepts this argument, I would conclude that Abbott is both persuasive and relevant to this case. Abbott addresses a consolidation request “through trial” under Illinois Supreme Court Rule 384.6 Plaintiffs correctly note that the Illinois rule differs from the language of California Code of Civil Procedure section 404, but still I would conclude that the Seventh Circuit’s reasoning is persuasive here. Similar to the Seventh Circuit in Abbott, we are examining a request for coordination or consolidation that lists certain goals that could only be accomplished through a joint trial. See Abbott, 698 F.3d at 573. As the Seventh Circuit did, we should have concluded that Plaintiffs were proposing a joint trial, and that federal jurisdiction under the CAFA mass action provision is proper.
In light of the specific reasons given for coordination of the California actions that involve propoxyphene, it is a natural and probable consequence of the grant of the petition seeking coordination, indeed it seems an inevitable result, that these varied actions must be tried together, or coordinated in a way to avoid inconsistent results as with bellwether trials, which amounts to the same thing. If the natural and probable consequence of coordination of separate actions has an impact indistinguishable from joint trial, then it is sensible to treat such a petition for coordination as a proposal for joint trial. I conclude that the circumstances presented here are a proposal for a joint trial within the meaning of what Congress said and intended in CAFA, and for that reason would reverse the district court’s order granting Plaintiffs’ motion to remand.7
.In the petition Plaintiffs asked for coordination of their lawsuits for reasons including concerns that there could be potential "duplicate and inconsistent rulings, orders, or judgments,” and that without coordination, "two or more separate courts ... may render different rulings on liability and other issues.” After this petition for coordination was filed, Teva removed the case to federal district court under CAFA’s mass action provision.
. The Seventh Circuit has held that CAFA "must be implemented according to its terms, rather than in a manner that disfavors removal of large-stakes, multistate class actions,” and I agree. Back Doctors Ltd. v. Metro. Prop. & Cas. Ins. Co., 637 F.3d 827, 830 (7th Cir.2011).
. The amicus curiae Chamber of Commerce of the U.S.A. and amicus curiae PhRMA want us to revisit Tanoh, to say that it has no vitality and that plaintiffs cannot structure *926their complaints to avoid federal jurisdiction in light of the purposes of CAFA to curb class action and mass action abuses that have occurred in state courts. Although this argument by the Chamber of Commerce has some weight, I agree with the majority that this argument misunderstands the power of a three judge panel, which may not overrule a prior circuit precedent.
. Amicus -curiae Washington Legal Foundation argues that "joint trial” includes cases resolved in conjunction with each other, relying on the dictionary definition of "joint” and the statute's plain language. This argument has some weight, and with the majority I would say that “joint trial” does not mean everyone sitting in the courtroom at the same time. Washington Legal Foundation also asserts that whenever the California coordination provision is invoked, that in itself will be enough to constitute a proposal for joint trial. I would not need to go so far to resolve this case because I rely in part on Plaintiffs' petition's explanation that there was concern to avoid inconsistent judgments, and because this case does not factually present as one where only coordination of pretrial matters was requested.
. I agree with Chief Judge Easterbrook of the Seventh Circuit that "[cjourts do not read statutes to make entire subsections vanish into the night.” Bullard v. Burlington N. Santa Fe Ry. Co., 535 F.3d 759, 762 (7th Cir.2008).
. Illinois Supreme Court Rule 384(a) says: “When civil actions involving one or more common questions of fact or law are pending in different judicial circuits, and the supreme court determines that consolidation would serve the convenience of the parties and witnesses and would promote the just and efficient conduct of such actions. The supreme court may ... transfer all such actions to one judicial circuit for consolidated pretrial, trial, or post-trial proceedings.”
. In light of what I would decide, I would not need to reach Defendants’ alternative arguments that federal subject-matter jurisdiction exists on other grounds.