Abstract One of the most important post-enactment questions with which courts are wrestling with regard to the Class Action Fairness Act, for cases affected by it, concerns the burden of jurisdictional proof. In the broadest sense, the issue is who bears the burden of proving the existence (or nonexistence) of the federal district court's subject matter jurisdiction under the new statute. Traditionally, the party who desires to maintain the suit in federal court always has had the obligation of demonstrating the court's authority to hear the case. Shifting some of the burden so that the other side has to prove that federal jurisdiction is lacking constitutes a sea-change of enormous proportions. Yet, as to one aspect of the burden of proof under CAFA, this is exactly what the vast majority of courts—and every circuit court to consider the question—has found Congress intended. It is already possible to see the impact that reallocation of some of the burden is having on class action litigation. For the party who does not want to litigate in federal court, bearing this new burden of jurisdictional proof has meant she has had a much harder time getting out of that forum. And, in light of other, recent decisions (both specific to CAFA and more broadly) that have heightened the importance of the pleading stage and of gathering adequate prefiling discovery, reallocation of the burden may soon become virtually dispositive in terms of forum selection outcomes.
In this paper, Professor Hoffman describes how a statute that is entirely silent on the question of jurisdictional proof under CAFA has come to spawn two different burden of proof debates, producing, most remarkably, two opposing answers. The author shows that the arguments credited by courts that have adjudged Congress intended to shift some of the burden of proof onto the party opposing federal jurisdiction rest on a number of highly suspect doctrinal and empirical assumptions. Against the prevailing view, Professor Hoffman argues that there are sound reasons to conclude that CAFA does not shift any of the burden of jurisdictional proof from the party who desires to maintain the suit in federal court.
Burdens of Jurisdictional Proof Lonny Sheinkopf Hoffman
This is a story only a lawyer could love. Having passed the most significant legislative reform of complex litigation in American history, advocates of the Class Action Fairness Act1 have now set their sights on more technical battles of statutory interpretation and application. Their ambitions span a broad range: from trying to stretch the boundaries of CAFA’s reach, in the first instance, to divining new principles in the conflict of laws which bear, ultimately, on the propriety of class certification.2 Nuanced and dry, and lacking all of the fanfare that accompanied the highly-charged and politicized debate over the law’s passage, these procedural questions of statutory meaning are being decided quietly, out of the headlines. Nevertheless, the stakes could not be higher. These interpretative contests are shaping CAFA’s full breadth and scope. They are the law’s bone and sinew.3 One of the most important post-enactment questions with regard to CAFA, for cases affected by it, concerns the burden of jurisdictional proof. In the broadest sense, the issue is who bears the burden of proving the existence (or nonexistence) of the federal district court’s subject matter jurisdiction under the new statute. Traditionally, the party who desires to maintain the suit in federal court always has had the obligation of demonstrating the court’s authority to hear the case.4 Shifting some of the burden so that the other side (which, as a practical matter, means shifting the burden to the plaintiff, since a challenge to the federal court’s authority occurs most often after the defendant has removed the case from state court and the plaintiff has asked that it be moved back) has to prove that federal jurisdiction is lacking constitutes a sea-change of enormous proportions. Yet, as to one aspect of the burden of jurisdictional proof, that is exactly what the vast majority of courts—and every circuit court to consider the question—has found Congress intended.5 This is the civil procedural equivalent of saying that criminal defendants are now guilty until proven innocent.
Congress, of course, cannot legislate all of the details that arise in the soft spots between the substantive provisions in the law. Indeed, it usually doesn’t even try, leaving the courts to flesh these out as live controversies come before them. This ought not to deceive us into thinking there are few practical consequences that flow out of the procedural details that develop around the substantive law. It is already possible to see the impact that reallocation of some of the burden is having on class action litigation. For the party who does not want to litigate in federal court, bearing the burden of jurisdictional proof has meant she has had a much harder time getting out of that forum.6 Indeed, in light of other, recent decisions (both specific to CAFA7 and more broadly8) that have heightened the importance of the pleading stage and of gathering adequate prefiling discovery, reallocation of some of the burden may soon become virtually dispositive in terms of forum selection outcomes. What this means is that so long as choice of forum matters9—a proposition that was, importantly, a foundational premise behind CAFA’s passage10—placement of the jurisdictional burden of proof will continue to be of enormous consequence to forum selection and, by extension, to case outcomes, as well.11 Yet, Congress said not one word in this detailed jurisdictional statute about who bears what burden before the court.
In this paper, I describe (in Part I) how a statute that is entirely silent on the question of jurisdictional proof under CAFA has come to spawn two different burden of proof debates, producing, most remarkably, two opposing answers. In Part II, I then consider all of the arguments credited by courts that have adjudged Congress intended to shift a part of the burden of proof onto the party opposing federal jurisdiction. I will endeavor to show that these arguments rest on a number of highly suspect doctrinal and empirical assumptions. Against the prevailing view, I argue that there are sound reasons to conclude that CAFA does not shift any of the burden of jurisdictional proof from the party who desires to maintain the suit in federal court.
I have two ultimate ambitions for this paper. In looking carefully at this vital question of statutory interpretation, my immediate aim is to offer a perspective different from what has become the established wisdom with regard to CAFA. By doing so, I hope to widen and improve upon existing thinking with regard to this specific question of statutory interpretation. More broadly, I also hope this study may offer some analytic clarity which can be drawn upon when future battles erupt over procedural questions of statutory meaning that lie, commonly and unavoidably, in the soft spots of the law where Congress rarely legislates.
I. CAFA’s Two Burden of Proof Questions
With CAFA’s passage, Congress expanded the original jurisdiction of the federal courts to now encompass class action lawsuits that, prior to the statute’s passage, were confined to state court. This legislative accomplishment was achieved through the introduction of two key statutory features: aggregation of the monetary value of claims and a minimal diversity of citizenship threshold. For any putative class action case, the claims of all class members are to be added together and, if the sum exceeds $5,000,000 exclusive of interest and costs and at least one member of the plaintiff class is a citizen of a state different from at least one defendant, then the federal court “shall have jurisdiction” by virtue of 28 U.S.C. §1332(d)(2).12 This means that a class action that satisfies these criteria can be originally brought in federal and, if filed in state court, can be brought into the federal system thanks to Congress’s passage of §1453, the corresponding removal provision.13 CAFA impacts most class suits that raise state law claims, but not all of them. Obviously, any suit involving less than the $5,000,000 floor would be outside the statute’s reach. Additionally, expansive as CAFA is, there are some provisions that limit the breadth of this expansion. We’ll return to these in a moment. It is enough to say now that whether one characterizes the elements necessary to satisfy jurisdiction (such as the aggregate amount in controversy requirement or that of minimal diversity between the parties) differently than those elements that limit jurisdiction turns out to matter a great deal in the debate over who bears what burden of jurisdictional proof.
A. The Threshold or Prima Facie Burden It is a venerable principle of federal jurisdictional law that the party seeking to maintain a suit in the federal forum (whether by virtue of instituting it himself or removing a case instituted by someone else in state court) bears the burden of proving the existence of the district court’s subject matter jurisdiction over the case.14 Nevertheless, a debate emerged in the early CAFA case law as to whether Congress in the statute shifted some or all of this initial burden.15 While a few district courts initially leaned in the other direction, the vast majority—and all of the circuit courts to consider the question—have held that the initial burden of demonstrating the existence of subject matter jurisdiction lies with the party that wants to be in federal court.16 Adhering to this traditional placement of the jurisdictional burden of proof, the courts have described the principle as “near canonical”17 and a “well-settled practice in removal actions.”18 In roundly rejecting the suggestion that Congress altered this traditional placement of the burden of proof on the party seeking the federal forum, most courts have considered it critical that the statute is silent on this point. Such silence, the courts say, is almost dispositive evidence of congressional intent. “[I]t would be thoroughly unsound for [the Court] to reject a longstanding rule absent an explicit directive from Congress,” one court intoned.19 “We presume that Congress, when it enacted CAFA, knew where the burden of proof had traditionally been placed. By its silence, we conclude that Congress chose not to alter that rule.”20 Indeed, courts also typically reject reliance on the statute’s legislative history to support the argument that Congress intended to shift the burden of proof on the threshold showing of jurisdiction to the plaintiff. In Brill v. Countrywide Home Loans, for instance, the plaintiff argued that a Senate Committee Report specifically stated that “the named plaintiff(s) should bear the burden of demonstrating that the removal was improvident.”21 The Seventh Circuit refused to accept this argument, observing that “[t]his passage does not concern any text in the bill that eventually became law. When a law sensibly could be read in multiple ways, legislative history may help a court understand which of these received the political branches’ imprimatur. But when the legislative history stands by itself, as a naked expression of “intent” unconnected to any enacted text, it has no more force than an opinion poll of legislators…”22
It bears emphasizing that this distinction between text and legislative history untethered to text that the courts have drawn with regard to the threshold jurisdictional showing under CAFA is not unique, of course, to CAFA. In this regard, the CAFA cases rejecting efforts to shift the initial burden of jurisdictional proof are straightforward applications of well-established general principles of statutory interpretation.23 If it has not been surprising, then, to see that most courts have refused to find that CAFA shifted the initial jurisdictional burden of proof, that makes stranger still the story of how a second burden of proof question, with regard to other provisions in CAFA that limit the scope of the statute’s reach, has come to be addressed and answered.
B. Shifting the Burden of Proof As To CAFA’s “Exceptions” While CAFA greatly expands federal jurisdiction, as we have seen, it also contains provisions that simultaneously limit the breadth of this expansion. The sentiment behind these parts of the statute ostensibly was to try to strike a balance between those cases that, in the Congress’s judgment, warrant federal jurisdiction and those that have “a truly local focus.”24 Oddly named, they are even harder to categorize. Section §1332(d)(4)(A) is regularly referred to as the “local controversy” exception, a phrase that does not appear in the statute, but is meant to get at the idea that if “more than two-thirds” of the putative class and at least one defendant is a forum citizen and alleged to bear “significant” responsibility for the plaintiffs’ alleged injuries, then the case ought to be litigated in state court since it is truly a local matter.25
A similar import lies behind §1332(d)(4)(B), which also is aimed at not allowing the statute to sweep local controversies into federal court, but has been dubbed, instead and rather clumsily, as the “home state” exception.26 This rule filters out just slightly more cases, in the sense that it applies where “two-thirds or more” of the putative class members are from forum, a figure which apparently but inexplicably includes those cases where exactly 66.6% of the class members are forum citizens (as opposed to the “local controversy” provision which requires that the figure be at least .01% more).27
At the same time, a second and more significant requirement of §1332(d)(4)(B), that “all” defendants must have been “primary” wrongdoers, seems intended to restrict the application of §1332(d)(4)(B) as compared with §1332(d)(4)(B). Whatever difference there is between a “primary” and “significant” wrongdoer (several courts have concluded that the Congress intended there to be a difference, though the practical distinction may be fairly small28), by requiring that every named defendant be a “primary” defendant, this statutory condition plainly is intended to more severely restrict the provision’s applicability.29 Still other parts of CAFA exclude from federal jurisdiction other kinds of cases, such as those where the defendant is a state, state official or other governmental entity,30 those that involve certain securities claims,31 and those cases raising state claims regarding the internal affairs or governance of a company or other business entity.32 In addition to and unlike all of these mandatory exclusions from jurisdiction, one other provision in CAFA gives district judges discretion to decide whether to keep a case over which CAFA has extended jurisdiction. The “interest of justice” exception, as it has been called by others, is found in §1332(d)(3), and covers cases where between one-third and two-thirds of the proposed class members and the primary defendants are forum citizens.33 How are all of these provisions in CAFA to be characterized? We’ve seen already that they are described by most commentators and courts as “exceptions” to jurisdiction, as in the “local controversy exception,” “home state exception,” “state action exception,” etc. Far and away, this is the predominant way the provisions have been understood, even though the word does not appear in any of these provisions.34 Nevertheless, they have come to be characterized by nearly all courts to consider them as “exceptions” to the expanded grant of jurisdiction in CAFA; specifically, that is, as exceptions to the expanded grant of jurisdiction in §1332(d)(2).
Calling them “exceptions,” those who hold to this view are distinguishing between jurisdiction-granting and jurisdiction-limiting rules. The notion goes something like this: If the elements that confer jurisdiction are satisfied, then the court has jurisdiction. Jurisdiction can be taken away by other provisions but—and this seems to be the important point—adherents of this view are certain that, at least in some theoretical sense, jurisdiction once conferred did exist, even if only for a fleeting moment.35 It can be undone, but its initial existence cannot be denied.
The notion that one can divine legislative intent by somehow distinguishing certain provisions as “exceptions” immediately raises a number of perplexing questions. For one thing, what textual proof is there that these provisions are “exceptions” to jurisdiction, in the sense that they are distinct from an a priori grant of jurisdiction elsewhere in the statute? Even if there were a plausible textual basis for distinguishing jurisdictional grants from exceptions in CAFA, why do the courts assume that all “exceptions” to jurisdiction in CAFA necessarily would have to be proven by the party who is arguing for their applicability? Outside of the specific CAFA context, the courts routinely require the party who desires to litigate in the federal forum to bear the burden of proof as to other, more familiar statutory provisions that limit removal rights. Consider, for instance, 28 U.S.C. §1441(b), the second sentence of which withholds in diversity cases a defendant’s right to remove that would otherwise exist under §1441(a) when at least one named defendant is a citizen of the state in which the action was filed. Though a limitation or “exception,” if you will, to the general removal right in §1441(a), the courts routinely place the burden on the defendant to prove that removal was not proscribed by §1441(b).36 Similarly, if the defendant’s removal is grounded on the assertion that the plaintiff fraudulently joined a non-diverse party solely for the purpose of destroying complete diversity, the removing party again has been held to bear the burden of jurisdictional proof as to the fraudulent joinder.37 The underlying rationale is the same in all of these instances: The party removing a case to federal court is obliged to demonstrate that the court has jurisdiction.
Still another perplexing question that is raised by the characterization of the CAFA provisions that limit jurisdiction as “exceptions” to jurisdiction is that, at best, this characterization seems appropriate only for the discretionary carve out in §1332(d)(3). If §1332(d)(3) were the only limitation of jurisdiction in CAFA, we would perhaps be able to say with more confidence that its language, which extends to the courts discretion to decline jurisdiction, makes it reasonable to regard it an “exception” to jurisdiction otherwise and previously given. But §1332(d)(3) is not the only provision that limits jurisdiction, as we have seen. Indeed, it is the only such provision of its kind, outnumbered and flanked on both sides by the mandatory exclusions from jurisdiction in §§1332(d)(4)(A), (d)(4)(B), (d)(5), (d)(9)(A), (d)(9)(B), and (d)(9)(C).
Along with these initial uncertainties that arise as to the prevailing view, I argue below that there are a number of other sound jurisprudential reasons for concluding, based on the statutory text, that Congress intended for the party who seeks the federal forum to bear the burden of proof as to §§1332(d)(3), (d)(4), (d)(5) and (d)(9). Yet, almost all courts—and every circuit court to consider the question—have held that Congress did intend to shift the burden as to §§1332(d)(3), (d)(4). Why?
We have already begun to hint at an answer: in the absence of express language from Congress, the characterization of the provisions as “exceptions” has been thought sufficient by the courts to justify shifting the burden of proof to demonstrate the applicability of these provisions. In this manner, courts have felt unconstrained by the traditional rule that speaks only to the burden of proving jurisdiction initially.
But why has this characterization of the provisions as “exceptions” been favored? And why are the courts so sure that, having decided that Congress meant these provisions to be “exceptions” to the grant of jurisdiction given in §1332(d)(2), Congress was signaling that the burden necessarily should shift to the plaintiff to prove their applicability? Most of all, how can we explain the startling level of agreement among courts in interpreting a statute that, on the specific issue of the burden of jurisdictional proof, is conspicuously silent and, from a broader perspective, has been characterized as “clumsy,”38 “bewildering”39 and “a headache to construe.”40 To get at the answers to these questions, it is necessary to take a much closer look at the leading cases and the arguments they have credited to justify this shift in the jurisdictional burden.
II. Behind the Cases A. The Assumption That the Plaintiff is in A Better Position to Access the Jurisdictional Facts One reason some courts have given to justify shifting the burden of proof in §§1332(d)(3) and (d)(4) is that the plaintiff is in the better position to access the facts supporting the existence or nonexistence of jurisdiction. Both the Eleventh Circuit in Evansand the Fifth Circuit in Frazier concluded that Congress intended to shift the burden of jurisdictional proof as to §§1332(d)(3) and (d)(4). Part of what led both courts to this judgment was their belief that plaintiffs generally have better access to facts relevant to the jurisdictional analysis called for by these statutory subsections. “[P]lacing the burden of proof on the plaintiff in this situation,” the Evans court intoned, “places the burden on the party most capable of bearing it.”41 The Fifth Circuit followed in stride.42 Oddly, both circuit court opinions talk in terms of plaintiffs, rather than the party who has invoked these subsections to try to get out of federal court. What the courts presumably mean by this is that since questions regarding §§1332(d)(3) and (d)(4) arise most often in the context of motions to remand, the plaintiff will normally be the party who bears the burden of jurisdictional proof. One charitably assumes that if a case were initiated in federal court and it was the defendant who was seeking dismissal, the Eleventh and Fifth Circuits would consistently leave the burden on the party who seeks to invoke §§1332(d)(3) and (d)(4).43 In any event, what about the empirical assertion made by these courts that the plaintiff—that is, the party seeking to establish that §1332(d)(3) or one of the two parts of (d)(4) applies—is in a better position to gather these facts? It is possible, of course, that the plaintiff sometimes will be in a better position to know particular jurisdictional facts, relative to the defendants, like the citizenship of members of the proposed class or the amount of each of the claims being sought. But it also seems pretty hard to believe that the plaintiff will always have better access to the relevant facts. Surely there will occasions—there were many in my own practice experience, primarily as a defense lawyer—when the defendant has a better vantage point. Indeed, it is hardly uncommon for the facts relevant to a claim to be more or exclusively in the defendant’s knowledge;44 it is not obvious why this would be different for jurisdictional facts. Notably, even courts that have ultimately shifted the burden of proof onto the party seeking to dismiss or remand the case from federal court have refused to go along with the Eleventh and Fifth Circuit’s reasoning here, rejecting the notion that one party necessarily and always has better access to jurisdictional facts.45