Takeaway: Class action complaints often seek the certification of “statewide” classes that would seem to include only residents of the target state. But when it comes to establishing an exception to diversity jurisdiction under the Class Action Fairness Act (“CAFA”), notably, the “local” or “home state” controversy exception, the evidentiary burden is on the class plaintiff. A district court’s determination that the exception applies has to be based on at least some facts in evidence. A recent decision by the Ninth Circuit Court of Appeals addressed the evidentiary requirements for proving up the exception, but while also noting that the burden of proof should not be all that difficult and emphasizing that a class plaintiff should be given the opportunity to conduct needed jurisdictional discovery.

The “local” or “home state” controversy exception to CAFA jurisdiction was the subject of the Ninth Circuit’s recent decision in King v. Great American Chicken Corp., Inc, — F.3d –, No. 18-55911, 2018 WL 4231847 (9th Cir. Sep. 6, 2018). The question presented was whether the class plaintiff had demonstrated that more than two-thirds of putative class members were California citizens at the time the case was removed to federal court under CAFA. And, more specifically, the Ninth Circuit panel had to navigate the sometimes fuzzy line between reasonable inference and proof by a preponderance of the evidence.

Plaintiff Celena King filed a putative class action in Los Angeles Superior Court on behalf of as many as 6000 non-exempt Great American Chicken (“GAC”) employees in California, asserting various violations of California wage-and-hour laws. GAC removed the case to U.S. district court (the Central District of California) pursuant to CAFA. King responded by propounding jurisdictional discovery in an effort to prove the local/home state controversy exception. Under that exception, when the primary defendant is a “local” citizen, a case removed under CAFA can be remanded if “two-thirds or more” of the proposed class members “are citizens of the State in which the action was originally filed.” 28 U.S.C. §§ 1332(d)(4)(A), (B). It was undisputed that defendant GAC was, in fact, a “local” citizen.

GAC resisted Ms. King’s jurisdictional discovery, contending that responses to her requests (seeking names, addresses, and telephone numbers), would not satisfy her burden to prove the necessary factual basis for remand. As an alternative, it offered to stipulate that “at least two-thirds [e.g., 67 percent]” of putative class members had last-known addresses in California. Ms. King declined this offer, but rather than decide the discovery dispute, the district court concluded – over GAC’s objection – that the stipulation was sufficient to meet the remand requirement. Accordingly, when King moved to remand, the district court granted her motion. The Ninth Circuit then granted GAC’s petition for permission to appeal the remand order.

Addressing the sufficiency of the evidence contained in GAC’s stipulation, the panel followed the Ninth Circuit’s ruling in Mondragon v. Capital One Auto Fin., 736 F.3d 880 (9th Cir. 2013). There, and consistent with decisions from the Fifth, Seventh, and Eleventh Circuits, the Ninth Circuit held that findings regarding class member citizenship had to rest on “at least some facts in evidence.” King, 2018 WL 4231847, at *3 (quoting Mondragon, 736 F.3d at 884). And whereas in Mondragon plaintiff’s reliance on a suggested class definition was deemed insufficient, even though reasonable inferences are permissible and a plaintiff’s burden of proof should not be “exceptionally difficult,” the same sort of deficiency was found to plague Ms. King’s position. Id.

Specifically, the panel observed that residency and citizenship are not the same. And even though GAC’s stipulation left some room for slippage in equating the two characteristics, the margin was too thin to justify an inference that the “two thirds or more” standard had been satisfied, because “at least” is not the same thing as “greater”: “[i]f we assume for the moment that the class included 6,000 members, as the district court estimated it might, two thirds would be 4,000, while 67 percent would be 4,020. ‘Greater than two-thirds’ would mean at least 4,001, so ‘at least 67 percent,’ or 4,020, would leave a cushion of only 19 class members.” Id. That cushion could too easily be consumed by the prospect that some class members had become citizens of other states before King filed her suit (as GAC had been able to show in its opposition to remand), by virtue of the fact that some last known addresses in GAC’s records were four years old or older when suit commenced, by the possibility that some GAC employees may have maintained citizenship in other states while working temporarily for GAC in California, and by virtue of the further possibility that some GAC employees might not even be citizens of the United States.

The panel did acknowledge that the district court’s inference was understandable, especially because jurisdictional discovery might become “burdensome and contentious.” Id. at *4. The panel, however, rejected the argument advanced by Ms. King that “at least two-thirds” could mean a figure substantially larger than 67 percent. The problem was that the argument “rest[ed] on guesswork.” Id. And, of course, guesswork is not enough. While a stipulation as to residency of putative class members cannot be rejected out of hand, a greater showing is necessary.

In light of these conclusions, the panel vacated the remand order and directed the district court to permit Ms. King to obtain additional jurisdictional discovery, especially given that the district court had denied her prior request for jurisdictional discovery. And, “[i]f GAC complains that the burden placed on it is too onerous, it is free to propose a stipulation that would better address King’s burden.” Id. at *5.

In the context of a wage and hour class action, satisfying the burden of establishing the local/in-state controversy exception to CAFA jurisdiction can be a difficult undertaking. Employers rarely if ever maintain information about the state citizenship of their employees. Statistically sound sampling and surveys, for example, might suffice, but they can be expensive. Of course, whether the effort to conduct discovery and develop this evidence is worth it will depend upon each individual case.