[PUBLISH]
IN THE UNITED STATES COURT OF APPEALS
FOR THE ELEVENTH CIRCUIT
________________________
No. 18-14797
________________________
D.C. Docket No. 8:18-cv-00381-WFJ-AAS
BETTY M. SMITH, as personal representative of the estate of Shirley T. Cox, JUDITH A. BALLEW, Attorney-in-Fact of John E. Ballew, MARK F. LAPP, as personal representative of the estate of Roger J. Lapp,
Plaintiffs-Appellees,
versus
MARCUS & MILLICHAP, INCORPORATED, Defendant,
MICHAEL BOKOR, Defendant-Appellant.
________________________
Appeal from the United States District Court for the Middle District of Florida
________________________
(March 12, 2021)
Before BRANCH, MARCUS, Circuit Judges, and HUCK,∗ District Judge.
BRANCH, Circuit Judge:
Three named plaintiffs, seeking to represent a putative class of 3,000 nursing facility residents, filed a class action complaint against Marcus & Millichap, Inc. (“MMI”), a real estate brokerage firm that marketed the relevant nursing facilities, and Michael Bokor, the president of the company responsible for managing the nursing facilities’ operations, in Florida state court. Bokor and MMI removed the case to the United States District Court for the Middle District of Florida pursuant to the Class Action Fairness Act (“CAFA”), which gives federal courts original jurisdiction over class actions where the amount in controversy exceeds $5,000,000 and there is minimal diversity between the parties (meaning at least one plaintiff and one defendant are from different states). 28 U.S.C. § 1332(d)(2).
But every statute has its exceptions. Here, the named plaintiffs sought remand to state court by invoking CAFA’s local controversy and discretionary exceptions. Those exceptions permit remand where a certain percentage of the
∗ Honorable Paul C. Huck, United States District Judge for the Southern District of Florida, sitting by designation.
3
putative class are citizens of the same state. See id. § 1332(d)(3), (4). To show that the proposed class met the exceptions’ citizenship requirements, plaintiffs provided the district court with twelve documents, including economic studies, statistics, and United States Census Bureau reports. They did not produce any evidence relating directly to the putative class, such as declarations of class members’ intent to remain in Florida, property records, or tax records. In this appeal, we consider whether these studies, surveys, and census data—which do not directly involve the plaintiffs in this case—are sufficient to establish that a certain percentage of the plaintiff class are citizens of a particular state for the purposes of CAFA’s local controversy and discretionary exceptions. We hold that they are not.
I. CAFA Jurisdiction
Before turning to the facts of this case, we begin with an overview of federal jurisdiction pursuant to CAFA. Through diversity jurisdiction, federal district courts have original jurisdiction over “all civil actions where the matter in controversy exceeds the sum or value of $75,000 . . . and is between . . . citizens of different states.” 28 U.S.C. § 1332(a)(1). In 2005, Congress enacted CAFA, which amended the federal diversity jurisdiction statute, 28 U.S.C. § 1332, to provide special rules for class action lawsuits. Class Action Fairness Act of 2005, Pub. L. No. 109–2 § 2(b), 119 Stat. 4. Pursuant to CAFA, federal courts have original jurisdiction over class actions where two conditions are met: the aggregate amount
4
in controversy exceeds $5 million, and the parties have minimal diversity— meaning at least one plaintiff is diverse from at least one defendant. 28 U.S.C. § 1332(d)(2), (6); see also Mississippi ex rel. Hood v. AU Optronics Corp., 571 U.S. 161, 165 (2014). To determine whether minimal diversity exists, courts consider the citizenship of all the class members (including putative), both named and unnamed. 28 U.S.C. § 1332(d)(1)(D). CAFA also includes specific provisions for the point in time when courts determine the plaintiff class members’ citizenship: (1) citizenship is first considered as of the filing date of the complaint or amended complaint; or (2) if the initial pleading does not state facts supporting federal jurisdiction, then citizenship is considered as of the date plaintiffs serve “an amended pleading, motion or other paper, indicating the existence of federal jurisdiction.” Id. § 1332(d)(7).
Several rules guide courts in determining parties’ citizenship. First and foremost, a natural person is a citizen of the state in which he is “domiciled.” McCormick v. Aderholt, 293 F.3d 1254, 1257 (11th Cir. 2002). “A person’s domicile is the place of ‘his true, fixed, and permanent home and principal establishment, and to which he has the intention of returning whenever he is absent therefrom[.]’” Id. at 1257–58 (quoting Mas v. Perry, 489 F.2d 1396, 1399 (5th Cir. 1974)). To put it another way, domicile (or citizenship) consists of two elements: residency in a state and intent to remain in that state. See Miss. Band of
5
Choctaw Indians v. Holyfield, 490 U.S. 30, 48 (1989). Residency is necessary, but insufficient, to establish citizenship in a state. Travaglio v. Am. Exp. Co., 735 F.3d 1266, 1269 (11th Cir. 2013). Courts look to various factors in determining a person’s intent to remain in a state, including: the location of real and personal property, business ownership, employment records, the location of bank accounts, payment of taxes, voter registration, vehicle registration, driver’s license, membership in local organizations, and sworn statements of intent. See, e.g., Sunseri v. Macro Cellular Partners, Ltd., 412 F.3d 1247, 1249 (11th Cir. 2005); McCormick, 293 F.3d at 1258; 13E Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 3612 (3d ed. 2019).
CAFA provides several exceptions to federal jurisdiction, pursuant to which a party can seek to remand the class action to state court. See 28 U.S.C. § 1332(d). Two of these exceptions—the local controversy exception, § 1332(d)(4)(A), and the discretionary exception, § 1332(d)(3)—are relevant to this appeal.
The local controversy exception provides that a “district court shall decline to exercise jurisdiction” over a class action that meets certain statutory criteria, two of which are relevant here. See id. § 1332(d)(4). The first addresses plaintiff citizenship: “greater than two-thirds of the members of all proposed plaintiff classes in the aggregate are citizens of the State in which the action was originally filed.” Id. § 1332(d)(4)(A)(i)(I). The second, referred to as the “significant
6
defendant test,” see Evans v. Walter Indus., Inc., 449 F.3d 1159, 1166 (11th Cir. 2006), requires the plaintiffs to prove that:
(II) at least 1 defendant is a defendant – (aa) from whom significant relief is sought by members of the plaintiff class;
(bb) whose alleged conduct forms a significant basis for the claims asserted by the proposed plaintiff class; and (cc) who is a citizen of the State in which the action was originally filed.
28 U.S.C. § 1332(d)(4)(A)(i)(II).
The discretionary exception provides that federal district courts may, “in the interests of justice and looking at the totality of the circumstances, decline to exercise jurisdiction” over a class action where: (1) “greater than one-third but less than two-thirds of the members of all proposed plaintiff classes in the aggregate” are citizens of the state in which the class action was originally filed; and (2) “the primary defendants are citizens of the State in which the [class] action was originally filed.” Id. § 1332(d)(3). If those threshold requirements are met, CAFA instructs courts to consider the following six factors:
(A) whether the claims asserted involve matters of national or interstate interest;
(B) whether the claims asserted will be governed by laws of the State in which the action was originally filed or by the laws of other States;
(C) whether the class action has been pleaded in a manner that seeks to avoid Federal jurisdiction;
(D) whether the action was brought in a forum with a distinct nexus to the class members, the alleged harm, or the defendants;
7
(E) whether the number of citizens of the State in which the action was originally filed in all proposed plaintiff classes in the aggregate is substantially larger than the number of citizens from any other State, and the citizenship of the other members of the proposed class is dispersed among a substantial number of States; and (F) whether, during the 3-year period preceding the filing of that class action, 1 or more other class actions asserting the same or similar claims on behalf of the same or other persons have been filed.
Id.
This case concerns the application of both exceptions.
II. Background
On January 5, 2018, Shirley Cox, John Ballew, and Mark Lapp (through their estates and personal representatives), individually and on behalf of all others similarly situated, filed a class action complaint against Bokor and MMI in Florida state court. Bokor is an individual Florida citizen and MMI is a Delaware corporation with its principal place of business in California. The complaint defined the class as “persons . . . who resided in any of the Facilities at any time during the period January 5, 2014 through [January 5, 2018].” In turn, the complaint defined the “Facilities” to include the 22 “skilled nursing facilities located throughout [Florida],” which MMI marketed and sold. To be eligible for admission to a skilled nursing facility in Florida, an individual must either (1) require long-term care because he cannot safely live alone in the home setting; or (2) require intensive rehabilitation before he can be safely discharged from the hospital.
8
Plaintiffs alleged that the facilities, including those where plaintiffs received care, improperly obtained licenses from the State of Florida by withholding critical information required by state regulations regarding their common ownership and management. Plaintiffs did not name the facility owners as defendants but contended that Bokor, who was the president of the company that operated the facilities and who personally submitted the facilities’ license applications, “used fraud and deception” to obtain the state licenses. They further alleged that, although MMI knew the licenses were invalid, the company endeavored to market and sell the facilities. As to harm that these alleged acts caused, the complaint alleged that “all of the residents at the Facilities during the relevant times were injured by being deceived into suffering substandard levels of care on a daily basis, which placed their health and well-being in jeopardy,” and that “each of the licensees of all the Facilities . . . shield[ed] the [operators of the facilities] from liability to future creditors, including Plaintiffs and the Class members.”
Plaintiffs brought five claims against Bokor and MMI: aiding and abetting breach of fiduciary duties (Counts I & II), civil conspiracy (Count III), and claims for civil remedies for criminal practices (Counts IV & V). They also sought actual damages for the class in the amount of $900,000,000, as well as treble damages, expenses, attorneys’ fees, and costs. The complaint estimated that the proposed
9
class of individuals “who resided in any of the Facilities at any time during the period January 5, 2014 through the date of this Complaint” exceeded 3,000 people.
On February 14, 2018, Bokor removed the action to the United States District Court for the Middle District of Florida pursuant to CAFA. See 28 U.S.C. §§ 1332(d)(2), 1453(b). Two days later, MMI joined Bokor’s notice of removal.1 Once in federal court, Bokor and MMI filed separate motions to dismiss plaintiffs’ complaint for lack of subject matter jurisdiction and failure to state a claim upon which relief can be granted. Plaintiffs did not respond to either motion. Instead, plaintiffs moved to remand to Florida state court pursuant to CAFA’s local controversy and discretionary exceptions. As discussed above, for either exception to apply, a certain percentage of the plaintiff class must be citizens of the state to which the case will be remanded. Accordingly, plaintiffs attached twelve items to their remand motion as proof that more than two-thirds (or, in the alternative, at least one-third) of the class were Florida citizens. These documents fell into three categories: (1) data generated by federal agencies, (2) economic studies concerning nursing facility markets, and (3) population migration surveys and reports.
1. Data Generated by Federal Agencies
1 The district court clarified in its remand order that “[t]he allegations of the complaint and notice of removal establish the prerequisites for removal under CAFA. Plaintiffs do not dispute Defendants have carried their burden to establish federal jurisdiction.”
10
The first document in this category is the Center for Medicare and Medicaid Studies “Nursing Home Data Compendium 2015 Edition.” Weighing in at 252 pages, the compendium includes population data not just from Florida, but also from forty-nine other states and the District of Columbia. While 96 pages explore nursing home resident characteristics, nothing in the document bears on where the nursing home residents lived prior to admittance or their citizenship. Further, the data concerning each state’s nursing home population is drawn from the 2010 census—which sampled the United States population as it was four years before the relevant time period in this case. In their motion to remand, Plaintiffs relied on this compendium for only two data points: (1) approximately 85 percent of nursing home residents nationwide are 65 or older; and (2) in 2013 and 2014, approximately 18 percent of all the nursing home residents in Florida had zero activities of daily living impairments. Notably, neither of those data points bears on of the citizenship of the putative class members in this case.
The next document in this category is the United States Census Bureau “Characteristics of the Group Quarters Population in the United States: 2016 American Community Survey: 1-Year Estimates.” This document shows that 85.4 percent of people in the United States currently live at the same address as they did one year prior and, of the 14.6 percent of people who moved in general, only 3.1 percent moved to a different state or abroad. And the survey data reflects the
11
national population—it does not address the state of Florida specifically or mention nursing facilities.
The final document in this group is the United States Census Bureau’s “2012-2016 American Community Survey 5-Year Estimates.” It presents Florida population data in a variety of fields, including household size, spoken languages, and income levels, but only devotes a few lines to “geographic mobility,” observing: (1)“[i]n 2012-2016, 84 percent of the people at least one year old living in Florida were living in the same residence one year earlier”; and (2) of the 16 percent that moved during that period, only 3.8 percent moved out of state. But it does not specifically address the geographic mobility of those 65 and older who lived in Florida during this period. And it does not mention nursing homes in Florida. Plaintiffs argued in their motion for remand that the two Census Bureau documents show that the plaintiff class members intend to remain in Florida.
2. Economic Papers from 2002 through 2015 Concerning Nursing Facility Markets
The second category consists of five economics-based academic studies and papers which address the geographic markets of nursing facilities.
The first is a 2015 working paper by Xin Zhao out of the University of Colorado, Boulder entitled “Competition, Information, and Quality: Evidence from Nursing Homes.” The author seeks to “estimate the effect of competition on nursing home quality and explore how the effect varies when consumers have
12
better access to quality information,” by extrapolating nationwide data. The paper includes data regarding the number of nursing homes in Florida from 2006 to 2011 but does not mention Florida anywhere else in its 47 pages. In their motion for remand, Plaintiffs rely on this document to support their argument that “an overwhelming majority of the proposed class members resided in Florida.” They pointed to only one line from this paper, which notes that “[t]he average distance between a nursing home and its potential consumer is 19.2 miles.”
The next document is a 2008 paper by David C. Grabowski, an associate professor of health economics in Harvard Medical School’s Department of Health Care Policy, entitled “The Market for Long-Term Care Services.” It aims to “highlight the measurement, data, and methodological issues underlying the study of long-term care markets and to suggest data that would improve our understanding of such markets.” The paper does not mention Florida or make any conclusions as to residency or citizenship. Plaintiffs rely on this document (along with the next three documents in this category) to support the proposition that “numerous peer reviewed studies have concluded to a statistically significant degree of certainty that eighty percent (80%) of the residents in any skilled nursing facility in any state were residing in the same county as the facility immediately prior to admission.” The paper does not quite support this conclusion. Instead, it reports merely that a 1994 study “found that 80% of residents in Wisconsin
13
facilities chose a nursing home located in the county where they resided before entering the home.”
Plaintiffs also included a 2002 study entitled “Use of Resident-Origin Data to Define Nursing Home Market Boundaries,” authored by Jack Zwanziger, Dana B. Mukamel, and Indridi Indridason from the University of Rochester School of Medicine and Dentistry. This study’s declared purpose is to test the assumption that “a nursing home’s market is coincident with the boundaries of the county in which it is located.” It relies on 1990 census data to examine Medicare beneficiaries admitted into a nursing home in New York state during the periods of 1992 to 1993 and 1996 to 1997 and finds that, for most New York nursing homes, more than 80 percent “of their market” resides within the county in which they are located.
Next up is a 2011 paper authored by John R. Bowblis, an assistant professor of Economics at Miami University and Phillip North, a portfolio analyst at Duke Energy, which is entitled “Geographic Market Definition: The Case of Medicare- Reimbursed Skilled Nursing Facility Care.” It evaluates “the different empirical techniques to define relevant geographic markets for nursing home care,” in order to “inform researchers and policymakers which method would be more appropriate given their research objectives and the constraints of their data.” This paper examines nursing facility markets in eight states, including Florida. It does not
14
provide any discussion or examination of Florida nursing facility residents, but includes a data point that 98.4 percent of Florida nursing home residents lived in Florida prior to entering the nursing home.
The final document in this category is a 2006 paper entitled “Spatial Competition and Market Definition in the Nursing Home Industry,” by Aditi Mehta—the paper does not say whether Mehta is a student or professor—in the Department of Economics of Boston University. The abstract for this paper illustrates the technical nature of the documents in this category:
The degree of competition in the nursing home industry has profound implications for regulations, quality and pricing decisions. This paper presents a model of spatial competition in the nursing home industry to investigate how consumer preferences over location can affect the substitutability and the degree of competition between geographically differentiated nursing homes. A random coefficients logit demand model for nursing home care is estimated using census tract demographic information. . . . The model is used to examine the impact of changes in the market structure of nursing homes in one hypothetical county.
The paper does not mention Florida or address citizenship of those in nursing facilities. Instead, it relies on 2002 CMS data concerning Wisconsin nursing homes to conclude that “the distance between the consumer’s previous residence and the nursing home is an important characteristic when choosing which nursing home to enter.”
3. Population Migration Surveys and Reports
15
Four surveys and reports make up this last category of documents. One document is a PowerPoint presentation, apparently created in 2011 by the North Carolina Division of Health Service Regulation, entitled “North Carolina Certificate of Need Law.” It examines health care facility services in North Carolina and does not mention Florida. As Plaintiffs emphasize in their remand motion, the presentation reflects that, in North Carolina at least, 74 percent of all hospital patients travel less than 25 miles to receive care.
Plaintiffs also relied on two surveys to show that the plaintiff class members intend to remain in Florida. One survey is the 45-page Pew Research Center report from 2008: “American Mobility: Who Moves? Who Stays Put? Where’s Home?” It surveys 2,260 people nationwide in 2008 regarding general moving trends across the United States, and indicates that 19 percent of individuals 65 or older “responded that [they were] very likely or somewhat likely [to] move in the next five years.” However, it provides no data specific to individuals in nursing facilities in Florida. Similarly, a 2012 survey titled “The United States of Aging,” which the National Council on Aging, USA Today, WPBT2, and United Healthcare sponsored, reported that “[c]lose to nine in 10 older Americans intend to continue living in their current homes for the next five to 10 years.” However, of the 2,250 people surveyed, only 250 were from Florida, and the survey says nothing about residents of nursing homes in Florida.
16
Finally, plaintiffs provided the district court with a 2014 article from a website belonging to what appears to be a business group named “Renaissance Planning” and written by Jeremy Goldstein, which they asserted helps establish the putative class members’ intent to remain in Florida. The Renaissance Planning article, entitled “Where Seniors Are Moving?” relies on the American Community Survey, discussed above, to point out that only approximately 10 percent of Americans 55 and older move. While it notes that Florida is one of several states that “had the most in-migration,” meaning individuals that moved to a new county or new state, “among people 55 and over” between 2010 and 2012, the article does not mention individuals in nursing facilities.
Bokor and MMI opposed the plaintiffs’ motion for remand on numerous grounds, arguing, in part, that the documents on which the plaintiffs relied were unauthenticated and they did not prove by a preponderance of the evidence that either CAFA exception applied.
While their motion for remand was pending, plaintiffs moved for leave to conduct discovery regarding the applicability of CAFA’s local controversy and discretionary exceptions to federal jurisdiction. Bokor and MMI opposed that discovery motion on the grounds that plaintiffs were merely seeking to bolster their argument for remand even though that motion remained unresolved.
17
On October 16, 2018, the district court granted plaintiffs’ motion to remand and denied plaintiffs’ discovery motion. The district court began its analysis with the local controversy exception. It divided the plaintiffs’ supporting documents into those purportedly showing each of the two separate requirements of citizenship for CAFA jurisdiction purposes: residency and intent to remain. As to residency, the court relied on the economic studies to show that “individuals who choose or land in a nursing facility hale from the proximate area,” and therefore “their residences were much more likely than not to be located in Florida.” The district court next examined the putative class’s “intent to remain” in Florida. In finding that two-thirds of the class intended to remain in Florida, the district court depended on federal census data, economic studies, and population surveys, which showed that United States citizens—particularly senior citizens—do not often move out of state. The district court concluded that:
Although . . . some of the studies may not focus specifically on Florida, the Court finds the data concerning nursing homes nationwide, together with the specific Florida studies, is sufficient to attribute citizenship to the putative class in Florida. Faced with the persuasive submissions of Plaintiffs, who carry the burden of proof of the exception to the CAFA, the Court finds that Plaintiffs have shown by a preponderance of the evidence that two-thirds of the class members are citizens of Florida.
And, after finding the other aspect of the exception satisfied—namely that Bokor was a significant defendant—the district court concluded that “[r]emand is warranted under the local controversy exception.”
18
Regarding CAFA’s discretionary exception, the district court further explained that,
To consider the discretionary exception . . . the class membership must be composed of greater than one-third but less than two-thirds of the class citizens of Florida. If the Court had not determined that two-
thirds of the class were Florida citizens, the Court would consider the six factors listed in the statute. To that end, the Court finds that this matter is purely local and not one of national importance. The claims will be governed by the laws of Florida and were not pled to avoid federal jurisdiction. There is a distinct nexus between this Florida action involving only Florida facilities, and the alleged torts committed and injuries sustained here. No other class actions asserting these claims have been brought in the preceding three years.
Bokor immediately filed a motion to stay the remand order pending appeal, which the court granted. Bokor then filed a notice of appeal of the district court’s remand order on November 15, 2018.
III. Standards of Review We review the district court’s decision to remand de novo. Evans, 449 F.3d at 1161. We review the district court’s evidentiary rulings for clear abuse of discretion. Aycock v. R.J. Reynolds Tobacco Co., 769 F.3d 1063, 1068 (11th Cir. 2014).
“When a party seeks to avail itself of an express statutory exception to federal jurisdiction granted under CAFA . . . the party seeking remand bears the burden of proof with regard to that exception.” Evans, 449 F.3d at 1164; see also Preston v. Tenet Healthsystem Mem’l Med. Ctr., Inc., 485 F.3d 804, 814 (5th Cir.
19
2007) (“Preston II”) (“[W]e employ the time-honored standard routinely applied to the fundamental question of citizenship” in determining whether plaintiffs meet a CAFA exception: “proof by a preponderance of the evidence.”). We review the district court’s jurisdictional factual findings concerning the parties’ citizenship for clear error. Travaglio, 735 F.3d at 1269. “A finding is clearly erroneous if the record lacks substantial evidence to support it.” Id.
IV. Discussion
A. The Local Controversy Exception We now determine whether the plaintiffs met the citizenship requirement under CAFA’s local controversy exception. That exception requires a district court to decline to exercise jurisdiction when three requirements are met: (1) greater than two-thirds of the proposed plaintiff class are citizens of the state of filing; (2) at least one “significant defendant” is a citizen of the state of filing; and (3) the principal injuries were incurred in the state of filing. 28 U.S.C. § 1332(d)(4)(A)(i).
Bokor claims that plaintiffs did not meet their burden as to the first two prongs of the local controversy exception and so the district court should not have remanded the case to state court. First, he claims that plaintiffs did not prove that at least two-thirds of the putative class were Florida citizens at the time they filed the claim in state court. Second, Bokor argues that the district court could not have
20
found that he is a significant defendant because plaintiffs did not present any evidence that Bokor could pay the judgment if he were to lose this case. We address each argument in turn.
1. Citizenship of Two-Thirds of the Plaintiff Class Bokor objects that the plaintiffs did not prove the first prong of establishing citizenship because even assuming that more than two-thirds of the putative class reside in Florida, plaintiffs did not submit any evidence that even one of the proposed class members intends to remain in Florida. Instead, according to Bokor, the U.S. Census data, unauthenticated private surveys, and studies cannot establish citizenship because they provide no evidentiary nexus to the putative class. 2 We
2 In determining whether the district court has jurisdiction over a removed case pursuant to 28 U.S.C. §§ 1332 and 1441, a federal court may consider “summary-judgment-typeevidence ”—meaning relevant evidence that would be admissible at trial. Williams v. Best Buy Co., 269 F.3d 1316, 1319 (11th Cir. 2001) (adopting the Fifth Circuit’s approach in Allen v. R & H Oil & Gas Co., 63 F.3d 1326, 1335–36 (5th Cir. 1995), and the Ninth Circuit’s approach in Singer v. State Farm Mut. Auto. Ins. Co., 116 F.3d 373, 337 (9th Cir. 1997)); Fed. R. Evid. 56(c). Nevertheless, evidence does not have to be authenticated or otherwise presented in an admissible form to be considered at the summary judgment stage, “as long as the evidence could ultimately be presented in an admissible form.” See Lossia v. Flagstar Bancorp, Inc., 895 F.3d 423, 429 (6th Cir. 2018); Maurer v. Indep. Town, 870 F.3d 380, 384 (5th Cir. 2017) (“At the summary judgment stage, evidence need not be authenticated or otherwise presented in an admissible form. . . . [Rather, the materials presented] need only be capable of being presented in a form that would be admissible in evidence.” (quotation omitted)).
Bokor makes much of the fact that the district court did not authenticate the plaintiffs’
documents when addressing the motion for remand, but as set forth above, at this preliminary stage, the evidence need not be authenticated to be considered—instead, it need only be capable of authentication. See Lossia, 895 F.3d at 429; Maurer, 870 F.3d at 384. Given the types of evidence presented—studies, census data, surveys, and reports—there is no indication that these materials would not have been capable of authentication later. See Fed. R. Evid. 901 & 902 (rules governing authentication of evidence). Accordingly, the district court did not abuse its discretion in considering this proffered data. See United States v. Siddiqui, 235 F.3d 1318, 1322
21
agree that the district court erred in finding this evidence sufficient to establish that two-thirds of the putative class were Florida citizens.
Class action plaintiffs can prove that two-thirds of the putative class are citizens of a certain state in two ways. The first way, as our sister circuits have recognized, is to limit the class definition to citizens of a certain state. See In re Hannaford Bros. Customer Data Sec. Breach Litig., 564 F.3d 75, 77, 81 (1st Cir. 2009) (recognizing that defining class to exclude “any persons and entities who are not citizens of the State of Florida” defeated federal jurisdiction under CAFA pursuant to the local controversy exception); Johnson v. Advance Am., 549 F.3d 932, 937–38 (4th Cir. 2008) (recognizing plaintiffs had “taken care” to avoid federal jurisdiction under CAFA by limiting the class to South Carolina citizens); see also In re Sprint Nextel Corp., 593 F.3d 669, 676 (7th Cir. 2010) (observing that if the plaintiffs had “defined their class as all Kansas citizens who purchased text messaging from Sprint Nextel or an alleged coconspirator” instead of Kansas residents, “the plaintiffs could have guaranteed that the suit would remain in state court”) (first emphasis in original).
Here, the class definition does not establish the citizenship of the class members for purposes of the first requirement of CAFA’s local controversy
(11th Cir. 2000) (“A district court has discretion to determine authenticity, and that determination should not be disturbed on appeal absent a showing that there is no competent evidence in the record to support it.”).
22
exception. Plaintiffs defined the class as individuals as “persons . . . who resided in any of the Facilities at any time during the period January 5, 2014 through [January 5, 2018.].” Again, residency does not equate to citizenship. Travaglio, 735 F.3d at 1269. Therefore, this class definition does not establish that the class is made up only of Florida citizens—meaning individuals who currently reside in Florida and have an intent to remain. Plaintiffs argue that this class definition helps establish the residency prong of citizenship, because it limits the class to Florida residents. But the definition does not limit the class to current Florida residents. Rather, it defines the class as persons who “resided” in the Florida facilities at some point during a four-year period. Thus, the class definition on its face encompasses class members who currently reside in the Florida facilities and those who resided in the facilities during the relevant four-year period, but have since moved to another state. Because citizenship for purposes of CAFA jurisdiction is based on current residency and an intent to remain, the class definition does not aid the class in establishing either prong of the citizenship requirement.
In cases where plaintiffs do not base citizenship on the class definition, they must provide evidence of the class members’ state of residence as well as evidence showing their intent to remain in that state. See Evans, 449 F.3d at 1165; see also Miss. Band of Choctaw Indians, 490 U.S. at 48 (“[D]omicile is established by
23
physical presence in a place in connection with a certain state of mind concerning one's intent to remain there.”). In determining someone’s intent to remain in a state, “[m]ere mental fixing of citizenship is not sufficient. What is in another man’s mind must be determined by what he does as well as by what he says.” Stine v. Moore, 213 F.2d 446, 448 (5th Cir. 1954). Thus, as we pointed out previously, courts look to various factors in determining a person’s intent to remain in a state, including property or business ownership, where the plaintiffs pay taxes and are registered to vote, and sworn statements of intent to remain. All of these forms of evidence relate directly to the parties asserting citizenship.
None of the evidence the plaintiffs provided is sufficient to establish either prong of citizenship. Instead, as we recounted above, the plaintiffs submitted only generalized studies and surveys—many of which date back to years before the relevant time period—and census data. While we do not hold that a district court may never consider evidence of a general nature in determining citizenship of the class, such generalized evidence cannot be the sole basis of the citizenship determination. Rather, in addition to any generalized evidence presented, “there must ordinarily be at least some [specific] facts in evidence from which the district court may make findings regarding class members’ citizenship for purposes of CAFA’s local controversy exception.” Mondragon v. Capital One Auto Fin., 736 F.3d 880, 884 (9th Cir. 2013); see also Myrick v. WellPoint, Inc., 764 F.3d 662,
24
665 (7th Cir. 2014) (rejecting plaintiffs’ reliance on a series of inferences based on the policy language to show citizenship for a class action policyholder suit against a health insurer, because the “plaintiffs did not offer any evidence to support” their propositions).
Turning to the case at hand, plaintiffs urge us to use “common sense” and make the following series of logical inferences from the generalized evidence they presented:
(1) Most people who go to nursing facilities resided in the state of the facility before admittance, so it is more likely than not that two-thirds of the putative class lived in Florida before going to the facility. 3
(2) Most people in nursing facilities in Florida are over 65, so it is more likely than not that two-thirds of the putative class is over 65.4
(3) Most people over 65 do not move out of state, so it is more likely than not that two-thirds of the putative class will not move out of Florida. 5
Plaintiffs assert that based on these inferences, it is more likely than not that two-thirds of the putative class are Florida citizens. 6 We cannot reach that conclusion here for several reasons.
3 The economic studies purportedly support this proposition.
4 The CMS Compendium purportedly supports this proposition.
5 The Census Bureau data, Pew Research Center report, and United States of Aging survey purportedly support this proposition.
6 We know that the named plaintiffs are citizens of Florida based on the exhibits attached to their complaint and the defendants’ concession below that each of the named plaintiffs were domiciled in Florida—i.e., resided in Florida and had the intention to remain there indefinitely. But the named plaintiffs’ citizenship does not establish the citizenship of the putative class.
25
First, none of the cited documents specifically addresses the citizenship (residency and intent to remain) of people admitted to Florida nursing facilities. In fact, only two documents specifically examine both Florida and nursing facilities. Furthermore, no document examines any resident of the 22 nursing facilities involved in this case. Second, the plaintiffs failed to define the type of skilled nursing homes at issue, meaning we do not know whether the facilities in question are long-term care facilities, short-term care facilities, or both—an important distinction that could affect the residency and intent to remain prongs of the citizenship inquiry. Third, while the evidence submitted tends to show that at least two-thirds of admissions to nursing facilities—for either long term or short-term care—generally are from the surrounding geographic area, the fact that a person lives in a certain geographic area does not establish necessarily that the person is a citizen of that state. That point is particularly true for a state like Florida where citizens of other states may live part of the year in Florida (and perhaps have to enter into a short-term care nursing facility while in Florida), but maintain a permanent residence elsewhere. In other words, without any specific evidence related to the putative class members, we cannot reasonably infer from the generalized evidence presented that the approximately 3,000 residents of the 22 nursing facilities at issue were Florida citizens. We cannot rely only on a series of purportedly reasonable inferences to determine citizenship; we cannot base our
26
determination of citizenship on “sensible guesswork.” Sprint, 593 F.3d at 674; see id. (“All in all, we’re inclined to think that at least two-thirds of those who have Kansas cell phone numbers and use Kansas mailing addresses for their cell phone bills are probably Kansas citizens. . . . But that’s all guesswork. Sensible guesswork, based on how the world works, but guesswork nonetheless.”); Brown v. Keene, 33 U.S. 112, 115 (1834) (“It is not sufficient that jurisdiction may be inferred argumentatively from its averments.”); King v. Great Am. Chicken Corp., Inc., 903 F.3d 875, 880 (9th Cir. 2018) (“The impression that this case would qualify for the local or home-state controversy exception is easy to understand. . . . The problem is that this impression rests on guesswork.”); Evans, 449 F.3d at 1166 (rejecting reliance on inferences to determine citizenship of putative class).
Moreover, even assuming that the plaintiffs had put forth sufficient evidence to show that two-thirds of the class members were Florida residents, none of the generalized data submitted was sufficient to establish the class members’ intent to remain in the state. At best, the submitted data speak only to population moving patterns and current residency, not to people’s intentions of moving. 7 If we
7 Plaintiffs claim that circumstantial evidence, like data and studies, is enough to show intent to remain. They point to discrimination lawsuits as an example of an area of law where courts use circumstantial evidence to show intent, like McDonnell Douglas Corp. v. Green, 411 U.S. 792, 805 (1973) (allowing evidence of “statistics as to petitioner’s employment policy and practice” to determine “whether petitioner’s refusal to rehire respondent . . . conformed to a general pattern of discrimination against blacks”) and Walker v. NationsBank of Fla. N.A., 53 F.3d 1548, 1555 (11th Cir. 1995) (recognizing the plaintiff could present statistical evidence to
27
accepted that current residency and statistics showing that less than 3 percent of people moved out of state between 2012 through 2016 proves that two-thirds of this class intends to remain in Florida, then we would be holding that proof of residency alone is enough to prove citizenship. Cf. Miss. Band of Choctaw Indians, 490 U.S. at 48; Travaglio, 735 F.3d at 1269; see also Hargett, 854 F.3d at 965 (“A complaint or notice of removal resting on residency, then, will not establish citizenship for diversity jurisdiction.”). Our precedent demands more.
Additionally, we have observed that the local controversy is a narrow exception, “with all doubts resolved in favor of exercising jurisdiction over the case.” Evans, 449 F.3d at 1163. With only generalized data and no specific facts to support the citizenship of any member of the putative class, doubts abound in this case. As the plaintiffs’ evidence fails to prove citizenship of any member of the class, it fails to establish more than two-thirds of the class are Florida citizens.
show discriminatory intent). In such cases, the plaintiff relied on statistics about the plaintiff’s particular employer to show a pattern of discrimination. But proving intentional discrimination is a different beast from proving intent to remain in a state for citizenship purposes. As discussed above, courts can look to evidence like voting and vehicle registration and property ownership to show intent to remain, and the plaintiffs themselves can sign an affidavit to that effect. In discrimination cases, however, evidence of an employer’s intent to discriminate is typically elusive. Regardless, even in those circumstantial evidence discrimination cases, the data must have some nexus to the parties. See, e.g., Pace v. S. Ry. Sys., 701 F.2d 1383, 1388 (11th Cir. 1983).
28
We therefore conclude that the district court clearly erred in its determination of state citizenship necessary for remand under the local controversy exception. 8 We recognize that class action lawsuits may become “totally unworkable” if the citizenship of each individual class member “had to be considered.” Preston II, 485 F.3d at 816 (quoting 1 Charles Alan Wright, Law of Federal Courts § 72, at 521 (5th ed. 1994)). And the “standard for establishing the domicile of more than one hundred plaintiffs must be based on practicality and reasonableness.” Id. But requiring something more than general data sources is not impractical or unreasonable.
2. Significant Defendant Bokor also argues that, in determining whether he was a significant defendant, the district court erred in not considering his ability to pay the potential judgment—plaintiffs are seeking more than $900 million in damages. He maintains that CAFA requires the district court to consider his finances in order to find him significant. We disagree.
As explained previously, in addition to the requirement that two-thirds of the proposed class be citizens of the same state, CAFA’s local controversy exception
8 But plaintiffs are not without a remedy. They could conduct a survey of the potential class. See, e.g., In re Sprint Nextel Corp., 593 F.3d at 675–76; Myrick, 764 F.3d at 665. And to prove their intent to remain, they could submit traditional forms of proof of intent for the court to consider, like the location of real and personal property, business ownership, employment records, payment of taxes, voter registration, vehicle registration, driver’s license, sworn statements of intent, etc. See, e.g., Sunseri, 412 F.3d at 1249.
29
requires plaintiffs to show that at least one defendant “from whom significant relief is sought” and “whose alleged conduct forms a significant basis for the claims” is a “citizen of the state in which the action was originally filed.” 28 U.S.C. § 1332(d)(4)(A)(i)(II). Such an individual is termed the “significant defendant.”
The district court found that Boker is a significant defendant for several reasons.9 First, the complaint seeks joint and several liability against both Bokor and MMI for all damages. Second, the district court determined that there was “no evidence that Bokor is any less or more significant with respect to liability . . . [n]or [was] there anything that would indicate that Bokor with his various entities is any less capable of paying a potential judgment than MMI.” Finally, the district court concluded that whether Bokor’s action forms a “‘significant basis’ . . . rests on the alleged scheme that contemplates each Defendant as equally culpable.”
Bokor does not dispute these findings. Rather, his only objection is that the district court failed to consider his ability to pay any potential judgment. In support of his argument that the district court was required to consider his ability to pay, Bokor relies on (1) our decision in Evans, (2) an unpublished district court opinion from the Western District of Louisiana, and (3) legislative history.10 To be
9 Because MMI is not a Florida citizen, the district court could not remand the action to Florida unless it determined Bokor, a Florida citizen, is a significant defendant.
10 We need not consider the unpublished district court opinion because that court’s opinions are not binding upon us. And legislative history is anathema to sound statutory analysis. See Conroy v. Aniskoff, 507 U.S. 511, 519 (1993) (Scalia, J., concurring) (“The
30
sure, the dicta in Evans stated: “whether a putative class seeks significant relief from an in-state defendant includes not only an assessment of how many members of the class were harmed by the defendant’s actions, but also a comparison of the relief sought between all defendants and each defendant’s ability to pay a potential judgment.” 449 F.3d at 1167 (quoting Robinson v. Cheetah Trans., No. Civ.A. 06- 0006, 2006 WL 468820, at *3 (W.D. La. Feb. 27, 2006)). But before we rely on dicta, we must turn to the statute itself.
We review de novo issues of statutory construction. Scimone v. Carnival Corp., 720 F.3d 876, 880 (11th Cir. 2013). “As in any case of statutory construction, our analysis begins with the ‘language of the statute.’ And where the statutory language provides a clear answer, it ends there as well.” Owens v. Samkle Auto. Inc., 425 F.3d 1318, 1321 (11th Cir. 2005) (quoting Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 438 (1999)).
The significant defendant provision is unambiguous in its requirement that the defendant be one “from whom significant relief is sought by members of the plaintiff class.” 28 U.S.C. § 1332(d)(4)(A)(i)(II)(aa) (emphasis added). That requirement does not equate to a defendant from whom significant relief may be
greatest defect of legislative history is its illegitimacy. We are governed by laws, not by the intentions of legislators. . . . If one were to search for an interpretive technique that, on the whole, was more likely to confuse than to clarify, one could hardly find a more promising candidate than legislative history.”).
31
obtained. See Coffey v. Freeport McMoran Copper & Gold, 581 F.3d 1240, 1244−45 (10th Cir. 2009) (“The statutory language is unambiguous, and a ‘defendant from whom significant relief is sought’ does not mean a ‘defendant from whom significant relief may be obtained.’”) Nothing in the statute indicates that district courts must conduct a factual inquiry into whether a defendant has the financial means to pay the damages alleged in the complaint. Thus, CAFA does not require the district court to examine a defendant’s ability to pay based on the unambiguous plain meaning of the statute’s text. And Bokor does not dispute that under this interpretation of the statute, he qualifies as a significant defendant. Thus, because we agree with the district court’s interpretation of the statute, we affirm the district court’s conclusion that plaintiffs satisfied the “significant defendant” requirement in § 1332(d)(4)(A)(i)(II)(aa).
Because we find that plaintiffs failed to meet the local controversy exception’s state citizenship requirement, however, the district court erred in remanding this matter to state court. B. Discretionary Exception To the extent that the district court also determined that CAFA’s discretionary exception applied, we now examine this exception.
As its name implies, the discretionary exception, unlike the local controversy exception, leaves the decision to remand largely up to the discretion of
32
the district court. However, in order to exercise that discretion, the district court must find two preliminary conditions are met: (1) greater than one-third but less than two-thirds of the aggregate members are citizens of the state in which the class action was originally filed; and (2) the primary defendants are citizens of the state in which the class action was originally filed. 28 U.S.C. § 1332(d)(3). Bokor argues that neither threshold requirement was met in this case. We agree.
1. Percentage of Florida Citizens The district court implicitly found that the plaintiffs met the first condition of the discretionary exception—that greater than one-third but less than two-thirds of the class members are Florida citizens—based on the same evidence it relied on in its analysis under the local controversy exception. As explained previously, our review of the evidence, however, determined that the plaintiffs’ evidence failed to prove the citizenship of any member of the class. Thus, that same evidence failed to establish that greater than one-third of the class are Florida citizens for purposes of the discretionary exception. See Preston II, 485 F.3d at 816 (“The same legal principles apply to the discretionary jurisdiction provision as apply to the local controversy . . . exception[]. Despite the burden to prove a lesser percentage of class members were citizens of Louisiana, which party bears the burden of proof and the sufficiency of evidence necessary to satisfy the citizenship requirements remains consistent throughout either analysis.”).
33
2. Primary Defendant Finally, Bokor contends that the discretionary exception is unavailable because MMI is a primary defendant but not a Florida citizen. The district court, however, did not address the discretionary exception’s requirement that all “primary defendants” are Florida citizens. 11 We do so now.
We recently grappled with what it means to be a “primary defendant” under CAFA in Hunter v. City of Montgomery, Alabama, 859 F.3d 1329 (11th Cir. 2017). We held that in deciding whether a defendant is a “primary defendant,” a court must ask “whether, given the claims asserted against the defendant, [the defendant] has potential exposure to a significant portion of the class and would sustain a substantial loss as compared to other defendants if found liable.” Id. (quoting Vodenichar v. Halcon Energy Props., Inc., 733 F.3d 497, 505–06 (3d Cir. 2013)).
Based on the allegations in plaintiffs’ complaint, MMI clearly is a primary defendant. Plaintiffs assert three causes of action against MMI—the same number
11 Instead of addressing this threshold primary defendant requirement, the district court proceeded straight to the statute’s six discretionary factors. It may have been the case that the district court assumed that the discretionary exception’s “primary defendant” analysis equates to the “significant defendant” analysis under the local controversy exception. It does not.
As explained previously, in order for the discretionary exception to apply, the district court must first find two preliminary conditions are met: (1) greater than one-third but less than two-thirds of the aggregate members are citizens of the state in which the class action was originally filed; and (2) the primary defendants are citizens of the state in which the class action was originally filed. See 28 U.S.C. § 1332(d)(3). Thus, both the percentage of Florida citizens and the primary defendant requirements must be found before the court can consider the six discretionary factors.
34
as they assert against Bokor. And plaintiffs seek to hold MMI and Bokor jointly and severally liable for more than $900 million in alleged damages: MMI has significant exposure and can potentially sustain at least the same loss as Bokor. But MMI is not a Florida citizen. MMI’s citizenship thus destroys plaintiffs’ ability to invoke the discretionary exception.
Plaintiffs put forth two arguments to lead us away from this conclusion.
First, plaintiffs counter that MMI is a corporation and therefore vicariously liable for the actions of employees, so it cannot be a primary defendant as compared to Bokor, who is directly liable for his own actions. Plaintiffs do not provide any authority for this argument and nothing in the statute or case law interpreting the statute indicates that courts should view individuals and corporations differently for purposes of determining primary defendants under CAFA’s discretionary exception. Second, plaintiffs argue that MMI is not a primary defendant because it did not steer this litigation; rather, MMI “simply joined” Bokor’s motion to remove and Bokor’s opposition to plaintiff’s motion to remand and only Bokor filed this appeal. Again, plaintiffs do not cite any authority to show why these facts should alter the “primary defendant” analysis above and the statute does not contemplate such an inquiry. Thus, to the extent that the remand order was based on the discretionary exception, the district court erred in failing to find that MMI is a primary defendant and not a Florida citizen.
35
REVERSED and REMANDED for further proceedings consistent with this opinion.