John Doe Defendants: Portents of Mystery, but Perhaps Not Diversity Jurisdiction

I. The Jurisdictional Problem of Anonymity

Imagine the following scenario. You and your friends have formed a band and recorded music, which you release on Spotify. Initially little more than a hobby based out of your parents’ garage in the suburbs of New York, the band quickly grows in popularity. In fact, several members of the band, yourself included, have started to rely on the income generated by your music, leaving behind other employment opportunities. Things are looking up.

In the habit of regularly googling your band name, you notice a series of negative, anonymous reviews. These reviews propagate, and soon, the demand for your music plummets. You and your friends decide to sue the owners of the anonymous accounts for defamation. From online searching, you are fairly positive you have identified one of the users—a man living in Florida—but you cannot determine the identity or location of the others, making them John Does. You file suit in federal court under diversity jurisdiction against the Floridian defendant and the John Does. To get their identities, you issue a pre-discovery subpoena to one of the John Does’ internet providers. The Florida defendant files a motion to dismiss under Federal Rule of Civil Procedure 12(b)(1)—asserting that the court lacks subject matter jurisdiction. Should the court grant it? More specifically, does having an anonymous John Doe as a defendant categorically preclude diversity jurisdiction?

Or, along other lines, imagine being a law student and seeing several anonymous accounts making posts both as you and about you on an online forum like Reddit. You have ascertained the identity of one account owner and are suing him plus the other account owners as John Does. The identified defendant moves to dismiss under 12(b)(1), too. Is diversity lacking?

These are not merely hypothetical circumstances; two district courts within the Second Circuit have handled cases with these fact patterns: Taylor v. Doe (S.D.N.Y 2021) and Doe v. Cioli (D. Conn. 2009). In both instances, the courts ultimately found that the parties, at the time, satisfied the diversity jurisdiction requirements, adopting what has been labelled a “wait-and-see” approach. Yet, had these cases been filed in Illinois and Wisconsin, instead of New York and Connecticut, courts would likely have dismissed the suits. Seventh Circuit precedent, namely Howell v. Tribune Entertainment Co. (7th Cir. 1997), holds that such defendants destroy complete diversity, unless the defendants are nominal, i.e., irrelevant to diversity, or when a named defendant wishes to remove to federal court. These divergent views split not only the circuits but courts within circuits, too.

In an era of ever-increasing technological reliance, anonymous harassment, defamation, and other offenses are bound to continue and will likely proliferate. Courts would do well to solve this divide. By allowing this genre of case to proceed under diversity jurisdiction, courts will save litigants and themselves time and effort, avoid arduous procedural gymnastics, and prevent serious jurisdictional harm.

II. The Basics of Federal Subject Matter Jurisdiction

There are, broadly speaking, two main paths by which federal courts may have subject matter jurisdiction over the average litigant: (1) federal question jurisdiction under 28 U.S.C. § 1331 and (2) diversity jurisdiction (for claims above the requisite amount in controversy) under 28 U.S.C. § 1332.1 The former path applies to all civil cases that arise “under the Constitution, laws, or treaties of the United States.” The latter—and the focus of this Essay—applies to cases between citizens of diverse citizenship where the matter in controversy exceeds $75,000.

To meet the diversity requirement, litigants generally must be from different states and/or countries. Since Strawbridge v. Curtiss (1806), federal courts have required complete diversity, meaning that no plaintiff and no defendant may be citizens of the same state. For example, a federal court would not have diversity jurisdiction over a case where a plaintiff from Massachusetts sued defendants from Vermont, New York, and Massachusetts.

III. John Does as Diversity Destroyers(?)

Disagreements among courts have left a question mark hovering over the ultimate rule for determining whether diversity jurisdiction exists in an action involving John Does. And there are few federal appellate decisions that answer the question, leading district courts within the same circuit to adopt varied approaches. In the Second Circuit, for instance, one court in the Southern District of New York held in Hai Yang Liu v. 88 Harborview Realty, LLC (2014) that diversity jurisdiction “cannot be invoked where unidentified Doe defendants, integral to determining diversity jurisdiction, are named in the complaint.” Meanwhile, the district courts in Cioli and Merrill Lynch Business Financial Services Inc. v. Heritage Packaging Corp. (E.D.N.Y 2007) allowed cases involving Doe defendants to proceed under diversity jurisdiction while reserving the right to dismiss at a later point should complete diversity be found lacking.

Though differing on the margins, the varied methods courts have taken can be broadly conceptualized as two approaches: (1) the Howell approach and (2) the wait-and-see approach.

In Howell, the Seventh Circuit held that anonymous defendants can be added in cases brought under federal question jurisdiction but not in cases brought under diversity jurisdiction. The line of reasoning was simple: a federal court ought not proceed without verifying that it has jurisdiction. Per § 1332, a court cannot know whether it has diversity jurisdiction unless the citizenship of all defendants is known or alleged. Therefore, in the absence of such knowledge or allegation, there can be no jurisdiction—with certain exceptions. Though not declaring its enumeration of these exceptions to be exhaustive, the Seventh Circuit’s opinion largely cabins the workarounds to three circumstances: (1) when the John Does are “merely nominal parties, irrelevant to diversity jurisdiction,” (2) in cases where a named defendant wishes to remove to federal court, and (3) in cases where a fugitive may have established a new domicile.

While the latter two exceptions are straightforward, the Seventh Circuit did not provide the clearest of tests as to what makes a defendant nominal. The court cited to Moore v. General Motors Pension Plans (7th Cir. 1996), where the Does at question, ultimately considered nominal, were placeholders for any additional parties that the plaintiff would add during or after the discovery process. Importantly, Moore distinguished these potential future defendants from situations where a plaintiff “knows that there are specific additional defendants he wishes to sue, but is simply uncertain as to their names.” Whether these types of defendants could ever be considered nominal remains hazy, with the Howell analysis seemingly saying no.

In subsequent years, as district courts around the country have faced this question, many have cited approvingly to the Howell approach, adopting similar or identical standards. For instance, in Stephens v. Halliburton Co. (N.D. Tex. 2003), Judge Sam Lindsay looked explicitly to the Seventh Circuit for guidance when addressing whether a plaintiff’s lawsuit against numerous defendants—including Does who were “past or present directors, officers, managing agents, and/or other employees or agents of [defendant corporation] Halliburton, whose identities are currently unknown”—defeated diversity jurisdiction. Judge Lindsay approvingly cited Howell and dismissed the suit for, among other things, lack of jurisdiction.

In Meng v. Schwartz (D.D.C. 2004), Judge Royce Lamberth held that the presence of John Doe defendants whose citizenships were not distinctly pleaded precluded diversity jurisdiction. Importantly, when addressing the plaintiff’s arguments that judges should assume plaintiffs would only ultimately add defendants that do not destroy diversity, Judge Lamberth, like Judge Lindsay, shot down such an approach, calling it “backdoor pleading . . . [that] require[s] the Court to draw the precise type of inference that the rules of diversity under § 1332 proscribe.”

In contrast, other courts (like the court in Cioli) have adopted a wait-and-see approach, allowing for diversity jurisdiction over Does until shown otherwise. If Howell represents one end of the spectrum of approaches to Doe defendant cases, MacHeras v. Center Art Galleries-Hawaii Inc. (D. Haw. 1991) serves roughly as the other end. In holding that Doe defendants should not defeat diversity jurisdiction, the MacHeras court relied heavily on equitable concerns in light of congressional direction on how to treat the effect of John Does on diversity jurisdiction in removal cases.

Some background is merited. According to the MacHeras court, the question of how to treat the effect of Doe defendants on diversity jurisdiction in removal cases plagued the Ninth Circuit in the 1980s, leading to a confusing series of tests and making the analysis a “near impossible determination.” To ameliorate the issue, the Ninth Circuit in Bryant v. Ford Motor Co. (9th Cir. 1987) decided to cut through the multifaceted standard and make a rule: “[T]he presence of Doe defendants under California Doe defendant law destroys diversity and, thus, precludes removal.” As such, the MacHeras court concluded, “no case could be removed to federal court so long as any Doe defendant remained a party.”

The Ninth Circuit’s decision faced backlash, leading to congressional action in the Judicial Improvements and Access to Justice Act of 1988. Therein, Congress amended 28 U.S.C. § 1441(b)—which governs the process for removing a case from state to federal court—and mandated that federal courts disregard “the citizenship of defendants sued under fictitious names” when determining whether diversity jurisdiction exists for purposes of removal.

Notably, both the Bryant decision and the corresponding federal legislation remained silent on diversity jurisdiction under 28 U.S.C. § 1332.2 While acknowledging this silence, the MacHeras court nonetheless found the logic of Bryant to be “equally applicable to cases brought in federal court under § 1332.” The court then assessed its options. To find, along Howell lines, that § 1332 precluded plaintiffs from including Doe defendants would have, in the court’s view, confined plaintiffs to state courts and unjustly granted anonymous defendants a unique forum selection advantage. To avoid this scenario, the court decided to grant jurisdiction.

Subsequent courts like the Southern District of New York in Weber Co. v. Kosack (S.D.N.Y 1997)3 , the Eastern District of New York in Merrill Lynch, and the District of Connecticut in Cioli have adopted glosses of this logic, turning it into the wait-and-see approach. Key to this approach is that the federal court can exercise jurisdiction over Doe defendants until it becomes aware of facts that would lead to dismissal. As each of these decisions emphasizes, plaintiffs bring these cases to federal court at their own peril, risking a sua sponte dismissal should one of the Does destroy complete diversity.

While these courts usually do not directly grapple with concerns that they may be exercising unconstitutional jurisdiction over the Doe defendants by exceeding the bounds set by congress, they do offer some potential rationales. For instance, in Cioli, the district court offered two reasons that legitimized its approach: first, that district courts have the authority to “cure a jurisdictional defect by dismissing a dispensable, nondiverse party,” and second, that appellate courts can restore complete diversity by dropping a party, even when the district court proceeded with improper jurisdiction.

IV. Policy Considerations and a Proposed Synthesis

The legal terrain generated by these cases presents a somewhat familiar scenario: a clash between what may textually be true and what may be most equitable. Precedent has firmly established that § 1332 requires complete diversity. And the Judicial Improvements and Access to Justice Act only amended § 1441, leaving § 1332 noticeably unchanged—at least regarding fictitious parties. Yet, to not allow filings against Doe defendants whose citizenships are unknown in federal court feels problematic. What would become of your aforementioned band? Or the law students? Since plaintiffs cannot remove to federal court, must they decide to either (1) forego the federal court route completely or (2) leave Does out of their federal complaint and file a separate, but virtually identical, suit in state court?

To mitigate these conflicting concerns, I propose that courts allow these cases to continue under diversity jurisdiction, but under close scrutiny. To that end, plaintiffs ought to be required to make a good faith effort to identify the citizenship of any Doe defendants. Courts can disagree as to what level of specificity this requires, but some amount of effort should be made. This approach is not entirely novel, having been selected as a middle ground between Howell and MacHeras by Judge Martha Vázquez in Hartford Casualty Insurance Co. v. Trinity Universal Insurance Co. of Kansas (D.N.M. 2015). There, the court held that the presence of Doe defendants does not destroy diversity as long as plaintiffs make a good faith effort to allege their citizenship. For example, in Hartford, plaintiffs alleged on “information and belief” that Doe defendants were not citizens of the respective plaintiffs’ states, a level of pleading the court ultimately allowed. Furthermore, it’s not clear that Howell precludes this reasoning, as that line of cases regularly dealt with claims against defendants where no allegations were made as to the Does’ citizenships.

That said, all involved parties, courts, plaintiffs, and defendants would benefit from a clearer definition of good faith. In Hartford Casualty, the plaintiffs based their allegations that the defendants were not citizens of any of the plaintiffs’ states upon “information and belief.” The court upheld this admittedly scarce explanation. What information and belief? In our band scenario, would an anonymous account incorporating a city or state name suffice? Would other references made by that account to geographic locations work? Or would information need to be more substantive, like a note from Reddit’s customer service team providing geographical information or an assessment of the account’s ISP? Certainly, that last example of jurisdictional evidence should suffice to keep a case in federal court, at least at the pleading stage. Beyond that, courts should ask whether a plaintiff presents enough evidence to plausibly plead diversity. The exact facts could—and should—vary from situation to situation.

Additionally, courts should allow plaintiffs to plead, as they did in Hartford Casualty, that Doe defendants are not from the plaintiffs’ state. In other words, our New York bandmates could get by, at least at the pleading stage, by either showing it likely that the Does live in a certain state that is not New York or by showing they have reasons to believe the Does do not live in New York.

While helpful, this still leaves open the question as to how courts should proceed if and when plaintiffs truly do not have a reasonable, good faith belief as to the defendant’s citizenship but nonetheless wish to sue in federal court. In these scenarios, federal courts should allow limited, and temporary, jurisdictional discovery. As it stands, jurisdictional discovery is not entirely foreign to federal courts. That said, its confines can be hazy and ill-defined, leading some to worry jurisdictional discovery will, in essence, allow for substantial and detailed proceedings before a court has established its jurisdiction, raising serious constitutional concerns.

In Jurisdictional Discovery in United States Federal Courts, Professor S.I. Strong addresses myriad concerns over jurisdictional discovery and proposes numerous paths forward. Outside of congressional action, one such path is judicial development of a hierarchy of facts designed to limit overreach. Questions of citizenship provide the perfect limiting factual framework. Courts should allow plaintiffs enough discovery to ascertain a reasonable belief as to the Doe defendant’s citizenship. Most simply, this could be a court issuing a subpoena to the website and/or internet provider to gain information on the defendant’s domicile. This need not—and should not—include any further information, not even a defendant’s name. Once ascertained, the plaintiff can decide to drop the defendant or continue, always bearing the risk that should evidence arise showing lack of diversity, the case will be dismissed.

Nonetheless, courts may still have constitutional concerns about proceeding with a case before jurisdiction has been conclusively established. These concerns are nothing to brush aside, but nonetheless do not require immediate dismissal on the pleadings. First, a good faith requirement and concomitant jurisdictional discovery procedures sufficiently protect the rights of the Doe defendants and should prevent any egregious constitutional violations. Even with the most passively invasive scenario, e.g., a court granting a discovery request as to a defendant’s citizenship, no substantive decision on the merits of the case has been reached.

Furthermore, allowing jurisdictional discovery promotes efficiency and, at worse, likely only shortens the timeline to discovery of the defendant’s domicile. In other words, were the case dismissed and the plaintiff forced to refile in state court, the plaintiff would likely be granted discovery there, yet could no longer switch to federal court. There, the defendants, if ultimately diverse, would have full power to decide whether to remove or not. The ultimate harm to the defendant, then, is merely the loss of this arguably inequitable right.

Finally, allowing these cases to proceed eliminates the odd statutory disjunction that the Judicial Improvements and Access to Justice Act created between § 1332 and § 1441. In fact, the statutory discrepancy that gives defendants permission to remove cases to federal court despite the presence of fictitious parties could lead to an unfortunate circularity. If there are both named and unnamed defendants in a suit filed in state court, the named defendant could remove the case to federal court. Once there, though, the unnamed defendants, the court sua sponte, or even the plaintiff could theoretically move to remand to state court, citing a lack of diversity jurisdiction. Theoretically, if there are numerous defendants, after remand another defendant could file a new notice of removal. Regardless of whether and when this circularity is cut off, its mere existence could generate significant procedural headaches for courts and litigants. Allowing for limited jurisdictional discovery could help break this procedural loop.

V. A Better Path for Future Litigation

Most of the stronger precedents allowing John Doe defendants, like Howell and Meng, predate social media. They were decided before the flourishing of online platforms like Twitter, Facebook, and the like. Given increasing anonymity in American society, some of the more recent decisions, which allowed cases to proceed with anonymous parties, provide a better path forward.

Adopting a clearer standard as to what a plaintiff must allege in federal court—a good faith, reasonable belief as to the Doe’s citizenship—largely maintains the status quo of having the burden of proving jurisdiction lie with the plaintiff. It simultaneously prevents what could amount to a de facto ban on plaintiffs suing anonymous defendants in federal court. At its most intrusive, the limited and targeted jurisdictional discovery needed for such an approach merely reveals citizenship along a faster timeline. If anything, requiring this identification process early on could help provide the defendants with notice, preventing further judicial decision-making—on either a state or federal level—without the defendants having the chance to participate. The standard also helps bolster the efficiency and efficacy of the federal courts in an ever-developing, increasingly tech-orientated world.

• 1There are, of course, other methods of gaining subject matter jurisdiction. Examples include cases where the United States is a party, cases involving ambassadors, and admiralty cases. See 28 U.S.C. §§ 1333, 1345, 1346 & 1351; see also William J. Hubbard, Civil Procedure: An Integrated Approach 866–67 (2021). This Essay focuses predominantly on diversity jurisdiction.
• 2The House Report is silent on the interplay with § 1332. H.R. Rep. No. 100-889, at 71–72 (1988).
• 3No. 96-cv-9581, 1997 WL 666246 (S.D.N.Y. Oct. 24, 1997).