Exhausting Government Class Actions

For decades, veterans were one of the few groups that could not bring class actions when the federal government systematically mishandled their disability benefits. In 1988, Congress passed the Veterans Judicial Review Act, creating a special “Veterans Court”—the U.S. Court of Appeals for Veterans Claims—that it expected would review veteran benefit decisions using class actions. But, just a year after the Veterans Court began hearing cases, that same court decided that it lacked the power to hear class actions in a short, summary opinion.

Without class actions, the Department of Veterans Affairs (VA) had persistently avoided judicial review of cases that impacted large groups of veterans. Facing seven-year wait times to wade through the VA’s complex benefit system, many unrepresented veterans gave up before ever seeing the Veterans Court. And, for the rare case challenging an unlawful rule or policy that actually reached the Veterans Court, judges observed that the government often strategically picked them off—quietly resolving petitions just before their hearing dates, ignoring the systemic problems alleged, and forcing courts to dismiss them as moot.

Five years ago, the Federal Circuit appeared to finally put an end to this. In a groundbreaking decision, it affirmed that the Veterans Court enjoyed the power to hear class actions like federal district courts that routinely review government agencies. First, it held that the Veterans Court could hear class actions under the All Writs Act—an infrequently used but important law that allows courts to craft novel procedures to protect their jurisdiction to decide cases. (This is the topic of my recent Article, The Class Appeal, appearing in the University of Chicago Law Review.) Second, the Federal Circuit said the Veterans Court could write a formal class action rule to hear class actions more regularly. The Veterans Court took this part of the decision to heart. After three years of meetings with government attorneys, veteran organizations, and academics, it created a formal rule to hear class actions for injunctive relief against the government and published it on Veterans Day, November 11, 2020. 

But just last month, what the Federal Circuit gave with one hand, it appeared to waive away with the other. In Skaar v. McDonough, the Federal Circuit held that the only veterans who could benefit from the new formal class action rules—at least those outside of the All Writs Act1 —were those veterans who had exhausted all of their internal appeals with the VA at the time that the class was certified. Given insurmountable backlogs and obstacles to legal representation inside the VA system, this holding could mean that few legal actions ever amass the numbers necessary to form a class in Veterans Court. Left unchanged by Congress or the courts, Skaar could leave in place the very administrative wrongs class actions hoped to repair—limited access to counsel, picked-off claims, and systemic problems—for thousands of veterans challenging the same unlawful policies at the VA.

As set out below, this result is not required by the Veterans Judicial Review Act or the basics of modern class action practice. In fact, the decision made two fundamental errors about the modern class action, which was specifically revised decades ago to clarify that class actions should benefit litigants even when some have not yet cleared every procedural hurdle imposed by a government agency. There is no reason that the Veterans Court cannot use those same common-sense tools to resolve recurring claims involving the nation’s veterans that federal courts have used to review other government programs for decades.

I. Veterans Class Actions and Skaar v. McDonough

A. The VA Benefit System and the Need for Class Actions

To understand how the Veterans Court came to embrace class actions, it’s important to understand four things about our nation’s veterans’ benefit system. First, the system is massive. Alongside immigration and social security, it is one of the largest mass adjudication programs in the United States. Over the past three years, it has processed three million claims for veterans seeking new or increased disability benefits because of an injury sustained in connection with their service in the military.

Second, the VA process is sprawling and Kafkaesque. After veterans apply for their benefits, they receive an initial administrative decision from one of 56 regional offices. Veterans who disagree with those initial decisions may request review by a senior adjudicator or request an appeal to the Board of Veterans’ Appeals, which in turn hears tens of thousands of claims inside the VA every year. Only when veterans have worked their way through all of those internal procedures can they finally appeal their claims to the Veterans Court—a special federal court in Washington, D.C., which hears more appeals from administrative agencies than almost all other federal appellate courts combined. At each stage of the benefits process, however, claims can be remanded for more factual development, often forcing claims to revolve up and down through the system on what some on the Veterans Court have called a “hamster wheel” of justice.

Third, the VA’s backlogs are intractable. Up until two years ago, the average benefit claim could take between five to seven years to be resolved. Congress has tried to modernize and streamline appeals. But problems in the VA’s new process have left veterans still waiting more than three years to fully resolve their claims.

Fourth, the VA system operates without meaningful precedent or access to counsel. The Veterans Court seldom issues precedential decisions. But even when it does, lawyers are rarely around to interpret and apply them consistently in the VA’s own proceedings. Low attorney fees, long delays, and ongoing remands make legal representation extremely challenging. Large volumes of unrepresented claims, in turn, aggravate inconsistencies, create more remands, and cause even more delays.

Against this backdrop, class actions offer an important tool to respond to recurring issues unrepresented veterans routinely experience at the VA. To be sure, many disability claims raise issues that are unique to a particular veteran. But when a group of veterans hopes to challenge the same illegal rule, arbitrary policy, or systemwide delay, class actions allow courts to ensure that veterans in the same boat receive the same relief before they get mired or lost in the system. In this way, class actions even help the VA—which doesn’t have to continually expend substantial resources hearing, remanding, and re-adjudicating the same common questions. As I explain in The Class Appeal, government mass adjudication programs benefit when reviewing courts can rely on class actions to aggregate claims and provide more consistent relief.

B. Skaar v. McDonough

A good example of how such a class action can work is the Skaar case itself. Skaar involved a class of nearly 1,400 elderly veterans. They alleged the VA arbitrarily denied them benefits, under the same flawed analysis, after they were exposed to dangerous ionizing radiation. The veterans in Skaar had been tasked with cleaning up nuclear debris after two Air Force planes carrying hydrogen bombs collided over Palomares, Spain in the 1960s. After the Air Force produced a study concluding that the cleanup did not expose the veterans to meaningful doses of radiation, the VA relied on the Air Force’s methodology to deny their applications for benefits. The class alleged that the standardized method that the Air Force and VA had used was scientifically unsound, and the Veterans Court eventually agreed. It certified a class of all veterans who “have been denied” and “will be denied” benefits under the program,2 ordered the VA to reconsider its methodology, and instructed the VA to adjudicate the Palomares claims accordingly.

In the process, the Veterans Court explained why the class action was particularly valuable for a case like this. Without a class, many of these older veterans would never know or be able to understand their right to receive a new hearing, obtain counsel, mount sophisticated challenges to a flawed scientific study, and obtain relief in a consistent way. The class action assured that—in one fell swoop—they would all get notice and new hearings and benefit from a new methodology with representation that they could not have obtained on their own.

The Federal Circuit reversed. It held that the only veterans who could participate and benefit from the classwide injunction were the few servicemembers who had already completed all their appeals inside the VA and were in a position to file with the Veterans Court at time the class was certified. It reasoned that the statute that gave the court jurisdiction to review the VA’s actions had to be read very narrowly. All the other veterans whose claims were still pending with the VA, whose claims had been wrongfully denied and not timely appealed, or whose claims had yet to be filed, were not entitled to the same relief. As a practical matter, this cramped view of the class action will mean that many veterans will not be able to mount class actions at all—especially given the obstacles to informing veterans about their rights to relief, the challenges of navigating the VA system without counsel, and the long delays they often would encounter raising the very same claims.

II. Skaar Ignores Foundational Assumptions Behind Class Actions

Class actions regularly proceed against government agencies in federal district courts, even those with the same kind of jurisdictional requirements at play in Skaar.3 This is, in part, because in injunctive relief cases, the named plaintiffs are “simply not asserting any claims that are not also applicable to” other members of the class. As long as the lead class representative jumps through any administrative hoops and jurisdictional prerequisites required by federal law, a court can assert jurisdiction over a class defined as those who “have filed” or “will file” claims for relief.

Nevertheless, the panel in Skaar distinguished those cases on two different grounds. First, it said that, unlike the Veterans Court, Congress gave federal district courts more expansive jurisdiction—under the federal supplemental jurisdiction statute—to hear claims that have yet to exhaust jurisdictional hurdles. Second, it said that defining the class to include those who “will file” claims with the court in the future is inconsistent with class action practice against the government. Both of these conclusions are wrong. They also misapprehend the purpose and historical rationale behind the modern class action rule.

A. Federal Jurisdiction to Hear Class Actions

First, the Skaar panel reasoned that Congress gave federal district courts more power to hear class actions than the Veterans Court. In so doing, it specifically pointed to the supplemental jurisdiction statute, which it believed gave federal district courts the power to hear claims brought by both those plaintiffs who met all the requirements of a jurisdictional statute and those plaintiffs who had not:

Mr. Skaar argues that the Veterans Court can exercise jurisdiction over class members who have not received Board decisions because district courts routinely certify classes including future claimants. Cross-Appellant’s Br. 26–30 (collecting cases). While district courts may indeed exercise jurisdiction over future claimants, that is because Congress explicitly conferred the district courts with supplemental jurisdiction encompassing such claims.

The panel’s analysis here was breezy and short, and perhaps it was confused by the arguments below. But it’s plainly incorrect. No federal court—not one—has ever said that the supplemental jurisdiction statute provides a basis to review federal class actions, asserting federal claims, against the federal government. And there are hundreds of class actions that have been asserted against the federal government, many with very similarly worded “jurisdictional” statutes as the Veterans Judicial Review Act.4

Why? Because that’s not what the supplemental jurisdiction statute is for. The supplemental jurisdiction statute allows federal courts to handle a mix of state law claims that they ordinarily cannot hear on their own under traditional federal question or diversity jurisdiction. But it was never meant to address federal claims against the federal government when some of those claims are not yet reviewable under another federal law. The text of the supplemental jurisdiction statute, which is missing from the Skaar opinion, makes that clear.5

Instead, when federal courts review a government agency, they rely on their interpretation of the federal question jurisdiction statute, as well as other statutes, to determine when to exercise that jurisdiction. When parties sue the Social Security Administration in a class action, for example, they sue under a court’s construction of 28 U.S.C. § 1331 and section 405(g) of the Social Security Act, which, like the Veterans Judicial Review Act, contains a jurisdictional prerequisite that claimants present claims to the Social Security Administration before they sue in federal court. The same is true of the Federal Torts Claim Act. Nothing about the supplemental jurisdiction statute enlarges or narrows federal district courts’ power to hear class actions against the federal government under a federal cause of action.

B. Satisfying Jurisdictional Statutes Using Class Actions

But the Federal Circuit’s mistake about supplemental jurisdiction belies an even larger one. The court appeared to hold that, in the absence of supplemental jurisdiction or some separate grant of authority, it was not enough for a class representative to finalize their claims at the VA before moving to certify a class in federal court. Instead, every class member must be able to establish the requirements for subject-matter jurisdiction in federal court at the time the class is certified. The government has made this argument in other cases in the past, but few courts have accepted it. Courts have rejected the argument because it’s inconsistent with why we have jurisdictional statutes, the law governing class actions, and the problems class actions historically set out to solve.

Start with jurisdictional statutes like the one that governs veterans’ claims: 32 U.S.C. § 7252 says that the Veterans Court will not have jurisdiction until veterans file their claims with the VA and are denied relief by the Board of Veterans’ Appeals, the final decision-making body at the VA. There are a number of reasons Congress would write a jurisdictional statute like this. First, it gives the government a first bite at solving the problem, thereby affording respect to a coordinate branch of government’s decision-making process. Second, the parties can fully develop the claim and record for a court to review. Third, jurisdictional limits save judicial resources; the court only has to devote its limited capacity to claims that demand an answer from a judge—which is particularly important in a mass adjudication system, like the VA, which hears over a million claims a year.

But nothing about a jurisdictional statute requires that a court give the government ten bites—or in a case like Skaar, 1,400 bites—at the same problem. Nor should it require the same expensive record development to repeatedly disprove the same flawed methodology, particularly given that many veterans will lack the access and resources necessary to develop that same record. Finally, reading a statute this way makes almost no sense in the context of a claim for injunctive relief; requiring the court to limit the scope of injunctive relief unnecessarily risks countless adjudications, over the same question, involving the same remedy, for the same scientifically unsound methodology. It would impose an arbitrary division between those whose claims were perfected at the time of an injunction and those whose claims were not yet ready. A hypothetical claimant who perfected a jurisdictional requirement the day before issuance of a classwide injunction might benefit from the judgment, while another claimant who did so the following day would not.

For that reason, the Federal Circuit in 2017 held that class actions were not inconsistent with the jurisdictional limits of section 7252.  The court even observed that congressional budget estimates released during the Veteran Judicial Review Act’s passage assumed that the Veterans Court would hear class actions over present and future beneficiaries’ claims.6

More pointedly, federal courts have long certified injunctive relief classes while making it clear that the class definition includes those who “have filed” and “will file” claims with a federal agency sufficient to satisfy the court’s jurisdiction.7 Those class actions strike the appropriate balance between the needs of representative litigation and the jurisdictional limits of our courts. Parties still participate in an administrative process that gives the government a chance to resolve an issue first. At the same time, courts can use class actions to apply the law without imposing an arbitrary limit on class membership (and relief) by excluding those with common claims who have not yet met certain procedural requirements at the earliest stages of litigation.

Even in strict statutory schemes, like the Social Security and Federal Torts Claims Acts, courts have long permitted class actions on behalf of members who “have or will” in the future satisfy jurisdictional requirements for the same reasons. To be sure, there has been some confusion on this question. The Skaar panel, for example, relied on a 1975 Supreme Court decision, Weinberger v. Salfi, that excluded absent social security class action members who had not yet presented claims to the Secretary under the strict language of the Social Security Act.

But shortly after Salfi, in Califano v. Yamasaki, the Supreme Court made it clear that class actions should not limit relief to only those people who satisfied jurisdiction at the time the class was certified. At issue in Califano was a flawed government process designed to claw back overpayments the government mistakenly made to Social Security recipients in the 1970s. The government originally allowed needy recipients to submit waivers to avoid the process, but only through written submissions and evidence. So, plaintiffs filed a class action, arguing that without live hearings and testimony, the waiver process violated Due Process and federal law. The Ninth Circuit generally agreed, affirming a class action that ordered the government to notify beneficiaries of their right to request waivers and receive live hearings. Beneficiaries were then given 30 days to exercise their rights to seek waivers in writing or in person after they received the government’s notice. 

The government appealed to the Supreme Court, arguing that the court could only assert jurisdiction and provide relief to class members who had already presented their waivers to the Social Security Administration when the class was first certified. The Supreme Court disagreed. It acknowledged that classes including “persons who had not filed requests for reconsideration or waiver in the past and would not do so in the future” were improper for certification because for those claimants “no ‘final decision’ concerning the right to a prerecoupment hearing has been or will be made” (emphasis added).  But a class action aimed at recipients who could claim a waiver would be certified because those people “are entitled to the opportunity for a hearing ‘when they claim a waiver’” (emphasis added) (quoting Elliott v. Weinberger, 564 F.2d 1219, 1222 (9th Cir. 1977)). Because the procedure for claiming a waiver involved “filing a written request with the Secretary,” the Court concluded, “we cannot agree that the Court of Appeals ordered this relief for those who do not meet the jurisdictional prerequisites of [the Social Security Act.]” In other words, it was proper for the district court to have ordered relief as to those people who had not yet filed waivers, as long as they would do so in the future consistent with a court order.

Since that time, the Supreme Court, as well as the lower federal courts, have had no problem with class definitions that expressly include those who may later be (but are not yet) in a position to benefit from the result of the class adjudication.8 To be clear, absent class members in such cases may still need to present their claims to the agency if they want individual relief—at least when a statute clearly requires that they do so. In other words, the certification of a class action does not alter the administrative exhaustion requirements or jurisdictional prerequisites that apply to individual determinations. But the class action procedure assures that when eligible recipients apply for relief, they will have meaningful access to the courts and lawyers who can assure the same, correct policy will be applied uniformly.

C. The Purpose and History of Classwide Injunctive Relief

This reading of class actions and jurisdictional prerequisites in the Veterans Court makes sense when one considers the broader rationale for class actions—particularly those that provide injunctive relief to a large group of people. Because class actions involve “representative litigation” over large, diffuse populations, courts generally have looked to the class representative to establish whether parties are diverse, whether venue is proper, and whether Article III subject-matter jurisdiction is satisfied. The rule that unnamed class members cannot defeat jurisdiction in such cases is justified by the goals of class action litigation. As the Supreme Court recognized in Devlin v. Scardelletti (2002): “Ease of administration of class actions would be compromised by having to consider the citizenship of all class members, many of whom may even be unknown, in determining jurisdiction.” If each class member had to separately satisfy a threshold procedural requirement before class certification, courts and administrative agencies would have to engage in the very sort of individualized, resource-intensive processes that aggregate litigation is designed to avoid. As a result, subject-matter jurisdiction usually exists as long as the named representative can establish it, the class definition is clear, and the class satisfies the procedural safeguards of Federal Rule of Civil Procedure 23.

Some exceptions have existed for diversity claims seeking money damages. For example, in the briefing in Skaar, the government relied on Zahn v. International Paper Co. (1973)—a diversity case that held that damage class actions required each class member to establish the amount in controversy—for the proposition that there is a “well established rule” that absent class members must separately establish subject-matter jurisdiction in class actions.9 In fact, most commentators considered Zahn the “outlier” in the world of class litigation, not a “well-established rule.”10 Importantly, even on its own terms, Zahn did not say when absent class members had to meet their jurisdictional requirements—which, as the Court clarified later in Califano, could take place when they applied to a government agency in the future.

Finally, the history behind the modern class action rule supports this approach to injunctive relief. Federal Rule of Civil Procedure 23(b)(2), the inspiration for the Veterans Court’s class action rule, was specifically revised in the 1960s to allow courts to review systemwide government misconduct, even when class members had yet to exhaust a lengthy administrative process. The effort to revise Rule 23 coincided with efforts after Brown v. Board of Education (1954) to desegregate public schools. By the early 1960s, a number of southern states had jettisoned crude, explicit policies that simply required segregated schools. Instead, school boards gave children a default school assignment but allowed them to individually petition school boards to have that assignment changed. Whether a board would grant any particular child’s petition ostensibly depended on a host of individual, facially nondiscriminatory factors specific to each one.

As administered, however, these processes left segregated schools almost entirely intact. Boards made default assignments by race, then systematically deployed a set of pretextual practices to reject individual petitions. When challenged in class actions, governments invoked these individualized government processes to argue that each student had to exhaust an individualized district application process. In a case called Potts v. Flax (5th Cir. 1963), the Fifth Circuit disagreed and affirmed the right to commence a class action: “Insofar as these statutes are advanced as prescribing statutory machinery which must first be administratively exhausted, we repeat what we have so often held. Exhaustion of internal school system administrative remedies is not required so long as racial segregation is the authoritative accepted policy.”

The Federal Rules Committee responsible for drafting Rule 23 relied extensively on Potts to highlight the importance of using class actions to join together parties who, although ostensibly in different positions, challenged the same government policy. Potts was included in the Advisory Committee’s note to Rule 23 as an exemplar of the Rule 23(b)(2) class action.

To this day, Federal courts have honored the Advisory Committee’s intention that Rule 23 provide a powerful tool to promote legal access, judicial review, and uniform relief, even when individual plaintiffs in a class were impacted in different ways at different stages of a governmental program. Veterans seeking classwide relief before the Veterans Court deserve no less.

III. Conclusion

The Class Appeal was inspired by the Veterans Court’s experiments with class actions. It concludes by reflecting more broadly about the hidden role technical procedures sometimes play in our separation of powers:

Abstract debates about the role of judicial review in our politics cannot take place without discussing the procedures courts use to perform it. Even for important courts that make binding decisions for large public institutions, individualized procedures can undermine their central role to hear claims, interpret law, and provide relief to our most vulnerable.

Although the focus in The Class Appeal sweeps more broadly to cover other complicated statutes that channel cases directly into our federal courts of appeals, the same lesson applies here. An overly rigid approach to jurisdiction in our public institutions can generate the same concerns as mandatory arbitration provisions that ban class actions against private institutions. Like individual arbitration, cramped interpretations of individualized exhaustion requirements before administrative agencies can also limit the judicial power to interpret law and prevent parties from pooling the resources they need to obtain legal representation and systemic relief against public institutions.

Class actions, however, have long been designed in ways that comply with the jurisdictional prerequisites set by Congress and permit courts to effectively manage the cases that come before them. In the process, as the Federal Circuit once observed, class actions may help the Veterans Court fulfill its own role “to compel correction of systemic error and to ensure that like veterans are treated alike.” A statutory scheme designed to assure benefits promised to “a special class of citizens, those who risked harm to serve and defend their country,” should not be undermined by narrow and mistaken readings of the law.  

• 1A separate rule was also created for class actions under the All Writs Act. See U.S. Vet. App. R. 22(a).
• 2The Veterans Court’s excluded from the class, however, veterans whose time to appeal inside the VA system had already expired under the VA’s old methodology.
• 3See, e.g., R.F.M. v. Nielsen, 365 F. Supp. 3d 350, 369 (S.D.N.Y. 2019) (collecting cases and explaining that under most statutory review schemes, the fact that the proposed “class includes members at various stages of administrative review does not defeat class certification”).
• 4See, e.g., Aiken v. Miller, 442 F. Supp. 628, 657–58 (E.D. Cal. 1977) (certifying a class of all those “whose application for food stamps was denied, delayed, or never made” and “who have been or will be affected by” the agency rule at issue); Lightfoot v. District of Columbia, No. 01-cv-1484, 2004 U.S. Dist. LEXIS 22158, at *8–12 (D.D.C. Jan. 14, 2004) (certifying a class of “[a]ll persons who have received or will receive disability compensation benefits . . . and whose benefits have been terminated, suspended or reduced” or “whose benefits will be terminated, suspended or reduced in the future” (emphasis added)); Barry v. Corrigan, 79 F. Supp. 3d 712, 751 (E.D. Mich. 2015) (certifying a class of “[a]ll past, present, and future applicants for, or recipients of, benefits administered by the Michigan Department of Human Services . . . who have suffered or will suffer actual or threatened denial, termination, or reduction of public assistance benefits”); Olenhouse v. Commodity Credit Corp., 136 F.R.D. 672, 681 (D. Kan. 1991) (certifying a class of farmers in an action to enforce their entitlements under federal grant-in-aid program and defining that class to include those who had not yet exhausted administrative appeals); Alexander v. Price, 275 F. Supp. 3d 313, 318 (D. Conn. 2017) (certifying a class of Medicare beneficiaries who “have received or will have received ‘observation services’”); Hill v. Sullivan, 125 F.R.D. 86, 87 (S.D.N.Y. 1989) (certifying a class of “widows or widowers who have or will apply for disability benefits”); McKenzie v. Heckler, 602 F. Supp. 1150, 1160 (D. Minn. 1985) (certifying a class of “all persons residing in Minnesota whose applications for SSI and RSDI have been or will be adjudicated concurrently”); Kendrick v. Sullivan, 784 F. Supp. 94, 104 (S.D.N.Y. 1992) (certifying a class of “all claimants for Social Security benefits whose claims have been or will be assigned to [the Administrative Law Judge] for decision”); Reed v. Lukhard, 591 F. Supp. 1247, 1251 (W.D. Va. 1984) (certifying a class of “persons in Virginia whose benefits . . . have been, continue to be, or will be denied, reduced, or terminated”); McDonald v. Heckler, 612 F. Supp. 293, 299 (D. Mass. 1985) (certifying a class of “[a]ll persons residing in Massachusetts who have filed or will file applications for disability benefits”); Newkirk v. Pierre, No. 19-cv-4283, 2020 WL 5035930, at *12 (E.D.N.Y. Aug. 26, 2020) (“[T]hat the class includes future members . . . does not pose an obstacle to certification.”).
• 5See 28 U.S.C. § 1367(a) (providing for jurisdiction over closely related claims except when “expressly provided otherwise by Federal statute”).
• 6See H.R.  Rep. No. 100-963, pt. 1, at 41–42 (1988) (discussing potential litigation challenges to VA regulations, stating, “[a]gain according to SSA, most challenges to regulations are class actions, involving large groups of beneficiaries or potential beneficiaries” (emphasis added)).
• 7See supra note 4 and accompanying text.
• 8See, e.g., Sullivan v. Zebley, 493 U.S. 521, 527 (1990) (affirming a class of children “who are now, or who in the future will be, entitled to an administrative determination” for children’s disability payments); J.D. v. Azar, 925 F.3d 1291, 1305 (D.C. Cir. 2019) (affirming certification of a class of pregnant children denied abortion access “who are or will be in the legal custody of the federal government”); Scott v. Quay, 338 F.R.D. 178, 192 (E.D.N.Y. May 25, 2021) (certifying a class comprised of “all those people . . . who have or will in the future have satisfied the exhaustion requirement imposed” by the Federal Torts Claims Act).
• 9Gov. Br. at 23.
• 10Of course, Congress made Zahn a dead letter when it passed the supplemental jurisdiction statute, as the Supreme Court laid out in Exxon Mobil Corp. v. Allapattah Servs., Inc. (2005).