Max W. Yarus 1 1 Max W. Yarus is an Appellate Litigation Attorney with Bergmann & Moore, LLC. B.A., Georgia College and State University; J.D., Stetson University College of Law.
- I. Introduction
- II. The CAVC’s Chosen Path: The Federal Rules for Civil Procedure
- III. Monk’s Definition of Commonality is Burdensome
- A. The Federal Circuit and VA’s low information nature require a more relaxed approach to certification
- B. Godsey and the Existing Path to Relaxing the Commonality Requirement
- IV. Class Modification in Practice
- V. Conclusion
Navigating new legal terrain, in the words of an en banc United States Court of Appeals for Veterans Claims (CAVC), is much like navigating
uncharted waters. 2 2 Monk v. Shulkin, No. 15-1280, 2018 WL 507445 at *2 (Vet. App. Jan. 23, 2018). In many ways, the CAVC is correct — one could imagine it traversing the reefs and corals of an unknown sea, running aground an unmarked shoal, or even implementing a historically rejected practice at the behest of a higher court. In Monk v. Shulkin, the United States Court of Appeals for the Federal Circuit reversed the CAVC’s decades-long practice of denying certification to classes of veterans. Class actions are formal aggregation procedures. 3 3 See Monk v. Shulkin, 855 F.3d 1312 (Fed. Cir. 2017); see generally Caroline Bressman, The Future of Class Actions, 104 Minn. L. Rev. 14 (2017); Michael Sant’Ambrogio & Adam S. Zimmerman, The Agency Class Action, 112 Colum. L. Rev. 1992 (2012). Now, just as plaintiffs may join a mass tort lawsuit, large groups of similarly situated veterans can collectively seek relief from the Veteran’s Affairs (VA) by alleging common legal and/or factual questions in a class format. The Monk remand was simple; the court held that the CAVC retained tripartite authority to entertain class action litigation but went no further as to decide on Monk’s merits or otherwise provide instructions to the CAVC on how to proceed in certifying classes. 4 4 Monk v. Shulkin, 855 F.3d 1312 (Fed. Cir. 2017).
At first glance, an instruction-less CAVC’s description of its seaworthiness with respect to an ocean of approaching class action litigation is apt. However, while
uncharted, the waters of claim aggregation in the veterans disability context are not completely unexplored. This article examines the recent class action litigation at the CAVC in Monk v. Shulkin, Godsey v. Wilkie, and Skaar v. Wilkie. 5 5 Godsey v. Wilkie, 31 Vet. App. 207 (2019); Skaar v. Wilkie, 31 Vet. App. 16 (2019). This article begins by identifying the rules the CAVC set for itself when certifying classes, then distills one of the major issues with sustaining veterans class actions — the commonality of the class’s members. After clarifying the definition of commonality as prescribed by the Supreme Court, this article uses the recent delay cases, Monk and Godsey, to show dual approaches to commonality at the CAVC. This article then proceeds to reconcile the CAVC’s two interpretations of commonality with contradictory precedent in Ebanks v. Shulkin and the VA’s low-information nature. 6 6 Ebanks v. Shulkin, 877 F.3d 1037, 1040 (Fed. Cir. 2017). Finally, this article contends that Skaar and other extant law provides a framework for relaxing the standard for commonality among veterans.
II. The CAVC’s Chosen Path: The Federal Rules for Civil Procedure
The CAVC is a creature of statute, born from a legislative act and given only the power to adjudicate claims under the jurisdiction granted to it by Congress. 7 7 Veterans Judicial Review Act of 1988, H.R. 5288, 100th Cong. (1988). Thus, the CAVC’s voyage into class actions began with caution, perhaps to avoid the appearance of an Article One judicial body expanding its jurisdiction on its own accord. In a January 2018 non-dispositive order — the CAVC’s first response to the Federal Circuit’s grant of authority — the CAVC stated that, while it was
considering adopting procedural rules to address class actions, it would initially forego independent rulemaking. 8 8 Monk v. Shulkin, No. 15-1280, 2018 WL 507445 at *2 (Vet. App. Jan. 23, 2018). Rather, the CAVC did as the Supreme Court often does by turning to the Federal Rules of Civil Procedure as needed. Here, turning to an appellate court was needed in order to fulfill the more trial-like duties of certification and the accompanying motion practice. The next time the CAVC addressed its framework for procedural class action rules was in an order denying certification of the Monk class. Again, the CAVC specifically declined to adhere to any concrete rules for certification, stating
[t]he Court anticipates that . . . it will adopt a rule on aggregate procedures that is appropriate for this Court. However, until that time, the Court will use Rule 23 of the Federal Rules of Civil Procedure as a guide. 9 9 Monk v. Wilkie, 30 Vet. App. 167, 170 (2018).
The CAVC’s preliminary decision to follow Rule 23 immensely benefits veterans seeking class-wide resolution. Foremost, it gives claimants freedom to assume that the final rule would not depart too drastically from the Federal Rules of Civil Procedure. Beyond that, the adoption of a framework reduces uncertainty. Having been employed by district courts for decades, the certification elements in Rule 23(a) are four-part and widely known: numerosity, typicality, adequacy, and commonality. 10 10 Compare U.S. Vet. App. R. 23, with Fed. R. Civ. P. 23. In the CAVC’s two certification decisions regarding delay in the VA, Monk and Godsey, the three former elements were dealt with swiftly. The
crux of the debate at the CAVC has instead centered primarily on diverging approaches to commonality.
A. Defining Commonality Broadly at the Supreme Court
The decision to adhere to the certification rules in the Federal Rules of Civil Procedure was not the only thing the CAVC borrowed from the Supreme Court. In both Monk and Godsey, although differing in their interpretations of the guiding precedent, the CAVC relied on the Supreme Court’s definition of commonality as interpreted by the 2011 decision in Wal-Mart v. Dukes. 11 11 See Monk v. Wilkie, 30 Vet. App. 167, 175 (2018); Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011).
Commonality in Rule 23(a)(2) requires that there be
questions of law or fact common to the class. 12 12 Fed. R. Civ. P. 23(a)(2).Although the rule seems straightforward, the Supreme Court found its plain language
easy to misread, since any competently crafted class complaint literally raises common questions. 13 13 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 349 (2011) (citing Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L. Rev. 97, 131–32 (2009)). In Wal-Mart, two types of common questions were identified: first being deep, probing questions, while the second being considered as
surface-level questions. Mere surface-level questions alleging commonality were found insufficient to satisfy Rule 23(a)(2) because, while they are important to individuals within the class, surface-level questions are not dispositive to the class as a whole. They do not implicate a deeper ultimate issue shared by each class member. Instead, to bind the class together, the Supreme Court required questions with depth, which identified a dispositive common injury and alluded to specific questions of law or fact. 14 14 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011). Put another way, common questions are the glue holding the class together. Like finding a weakness in one of the stitches in a ship’s sails would spell disaster, so too would finding a class question that fails to deeply implicate a dispositive issue.
For example, the plaintiffs in Wal-Mart claimed a violation of Title VII of the Civil Rights Act. The complaint alleged a company-wide practice of discrimination in millions of hiring and payment decisions against women and that a strong corporate culture in favor of men essentially made
every woman at the company the victim of one common discriminatory practice. 15 15 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 345 (2011). Notably, employment law dictates that the ultimate issue shared by all Title VII claims is the reason for the employment decision. One could then imagine, as the Supreme Court did, questions that are common to the putative class, but which do not generate common answers to the ultimate merits question, include:
Do all of us plaintiffs indeed work for Wal-Mart? Do our managers have discretion over pay? Is that an unlawful employment practice? What remedies should we get? It is important to note that no question quoted above highlights a specific common injury to the millions of putative class members. Nor does any question overlap with the ultimate merits inquiry. A question sufficient to bind the whole class might then be one alleging that the same employment practices
touch and concern each class member, such as
the assertion of discriminatory bias against a specific supervisor. 16 16 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 349–50 (2011). This question shows that the key to commonality is not the question or answer itself, but the capacity for a common question to generate a common answer to a dispositive issue. Questions to that effect will generate dispositive proof for each class member and limit who can be common to that class. In this way,
proof of commonality necessarily overlaps with [a classes’] merits contention, and the practical consequences of considering that proof would be the disposition of each class member’s claim, regardless of outcome. 17 17 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 352 (2011) (citing Cooper v. Fed. Reserve Bank of Richmond, 467 U.S. 867, 876 (1984)).
Finding that common question difficult, the Supreme Court ultimately approached this task by weighing the class members’ differences against each other. A question that allows too many class members with distinct legal claims or factual inconsistencies would not satisfy commonality, but a question that implicates a shared dispositive inquiry would:
What matters to class certification . . . is not the raising of commonquestions— even in droves — but, rather the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation. Dissimilarities within the proposed class are what have the potential to impede the generation of common answers. 18 18 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011) (quoting Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L. Rev. 97, 132 (2009)).
Thus, the only way to bind a class under Rule 23(a)(2) is to find a positive balance between common questions that generate common answers and the dissimilarities of a class. Because legal distinctions erode commonality, probing the putative class for dissimilarities is a crucial last step to sustaining aggregation.
After identifying common questions, the Wal-Mart Court did just that, and considered the degree to which the differences among class members would impede the generation of common answers. Finding that the complaint stretched amongst
3,400 stores, sprinkled across 50 states, with a kaleidoscope of supervisors . . . subject to a variety of regional policies that all differed, the Supreme Court denied the Wal-Mart class because too many dissimilarities existed to efficiently litigate the class’s claims as a unit. 19 19 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 359–60 (2011) (citing Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571, 652 (9th Cir. 2010) (Kozinski, C.J., dissenting)). If the class had instead proposed to bind themselves by alleging discrimination as the reason for a specific supervisor’s employment decisions, then asking this single question would eliminate dissimilarities among a potential class because its answer could prove whether or not each class member suffered the same injury.
An important ambiguity in Wal-Mart is the majority’s apparent conflation of Rule 23(a)(2) with Rule 23(b)(3)’s predominance requirement. Rule 23(b)(3) states that a class action for money damages can be maintained when class questions predominate over questions affecting individual members. 20 20 Fed. R. Civ. P. 23(b)(3). The concurrence in Wal-Mart argues that the majority’s focus on dissimilarities forces the decision-maker into a predominance analysis under Rule 23(b)(3), improperly heightening the burden a class bears to meet the commonality requirement. 21 21 Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 376–78 (2011) (Ginsburg, J., with Breyer, Sotomayor, and Kagan, JJ., concurring in part and dissenting in part). The concurrence neglects to consider the practical effect of a dissimilarities analysis, which is an effort to determine the existence of a sufficiently common question by weeding out legal and factual distinctions. As the majority notes, such a
rigorous analysis is proper in light of the fact that
Rule 23 does not set forth a mere pleading standard. 22 22 See Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350–51 (2011) (citing Gen. Tel. Co. of Southwest v. Falcon, 457 U.S. 147, 160–61 (1982). Considering the Supreme Court concurrently maintained that even a single adequate question has the power to bind a class and sustain aggregation, the standard for commonality remains low even outside of veterans law.
III. Monk’s Definition of Commonality is Burdensome
The debate highlighted in Wal-Mart recently departed from the tort context and entered the seas of veterans law. Whereas in employment law, precedent has long established that the root of a Title VII claim is
the reason for a particular employment decision, no such precedent yet identifies a controlling question for aggregate veterans’ issues. 23 23 See Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 352 (2011) (quoting Cooper v. Fed. Reserve Bank of Richmond, 467 U.S. 867, 876 (1984). The CAVC has considered issues of commonality multiple times since assuming the authority to entertain class actions, first introducing a strict standard in Monk and then relaxing it in Godsey.
The Monk class was the first considered by the CAVC for certification. By way of background, prior to the enactment of the Veterans Appeals Improvement and Modernization Act of 2017, typical disability claims proceeded through the VA on a single track regardless of the claim’s details. 24 24 See Veterans Appeals Improvement and Modernization Act of 2017, Pub. L. 115–55, 115th Cong. (2017). If a veteran received an unfavorable initial decision at the VA Regional Office (VARO), then he or she could file a Notice of Disagreement (NOD) to launch an inner-agency appeal. If a second review yielded another unfavorable decision, then the veteran could file a VA Form 9 to appeal to the Board of Veterans Appeals (the
Board). Before the Board heard an appeal, it must be certified by a higher-level review. If no further development of the claim is necessary, the Board would review the matter de novo.
The veterans in Monk argued that the total time the VA took to decide their appeals to the Board after submitting an NOD was unconstitutionally long and effectively deprived the veterans of the benefits sought. 25 25 Monk v. Wilkie, 30 Vet. App. 167, 169 (2018). The class contended that there was a maximum time to wait for a decision, beyond which the wait became unreasonable. They did not identify a practice or procedure within the VA that adversely affected each of their claims, but attempted to glue themselves together with two prevailing questions:
(1) Whether extensive delays in failing to render decisions on disability claims within 12 months of timely NODs violate the proposed class members’ due process rights; and (2) whether the proposed class members are entitled to a writ of mandamus compelling the Secretary to correct the severe delays and inaction under Telecommunications Research & Action Center v. FCC. 26 26 Monk v. Wilkie, 30 Vet. App. 167, 175–77 (2018) (citation omitted).
Like the surface-level questions considered by the Supreme Court in Wal-Mart, the CAVC found that neither proposed question generated answers common to the entire class.
Question one identified a claim that the government violated the procedural due process rights of veterans. In this constitutional challenge, the Monk plurality performed a dissimilarities analysis and reasoned that it would be critical to consider both the degree of constitutionality of the delay in each petitioner’s claim as well as the VA’s justification for the delay. The various reasons that could be implicated reveal class relief is impracticable in light of the need for an individual inquiry. To reconcile potential dissimilarities, instead of ordering mandamus relief for the class, the CAVC required a more individualized balancing test — such as the Mathews test — to assess the quality of the procedural due process afforded to each veteran. 27 27 Monk v. Wilkie, 30 Vet. App. 167, 186 (2018) (Allen, J., with Bartley and Toth, JJ., concurring in part and dissenting in part) (citing Mathews v. Eldridge, 424 U.S. 319 (1976)).
Question two identifies a claim that the delay in veterans’ appeals is unreasonable despite not amounting to a constitutional violation. This time, consideration of the six-factor TRAC test led the Monk court to the same conclusion as in question one: A class-wide writ of mandamus was not an efficient way to adjudicate thousands of delay cases based only on systemic delay. Like question one, the Monk court sought individual inquiries into the class members’ appeals by examining the reasonableness of the agency’s action. Because
more complex and substantive agency actions may be expected to take longer than
purely ministerial [or administrative] ones, and because each member may have suffered a different injury or reason for delay, another individualized inquiry is necessary. 28 28 Monk v. Wilkie, 30 Vet. App. 167, 177 (2018).
In finding the proposed questions inadequate, the plurality reasoned that delay claims can be aggregated only where the whole class suffered adversely from a specific policy or practice of the VA. The plurality further stated that
the petitioners [must] identify the reasons for delay . . . so that we may determine whether commonality exists. 29 29 Monk v. Wilkie, 30 Vet. App. 167, 178 (2018). Here, the CAVC relied on a Ninth Circuit case, Parsons v. Ryan. 30 30 Parsons v. Ryan, 754 F.3d 657 (9th Cir. 2014). In Parsons, a class of prisoners sought injunctive relief against the Arizona Department of Corrections (ADC) for deficiencies in the state prison healthcare system. What differentiated the certified class in Parsons from the putative class in Monk, was this issue of identifying a specific policy or practice to use as glue to bind the class. The Parsons class enumerated
17 statewide ADC practices that
exposed all inmates in ADC custody to a substantial risk of serious harm. 31 31 Parsons v. Ryan, 754 F.3d 657, 673, 676 (9th Cir. 2014). On the other hand, the Monk class failed to identify any policy or practice affecting the whole class, and even conceded in oral argument that no such common challenge existed.
Essentially, the plurality in Monk juxtaposed its own putative class with the Wal-Mart and Parsons classes. More like the class of women in Wal-Mart than the class of inmates in Parsons, the Monk class was found to be amorphous and filled with dissimilarities. The veterans in the putative class suffered from a wide range of disabilities. They also suffered from numerous reasons for delay, but they argued in good faith that the variance was irrelevant
so long as [each member’s] wait is more than 12 months. 32 32 Monk v. Wilkie, 30 Vet. App. 167, 175 (2018). However, a plurality of the CAVC was not persuaded by the veterans’ argument that the varying reasons for delay were unimportant. Instead, the Monk court introduced a strict requirement on veterans to affirmatively know the reason for their delay, especially in light of a 2017 Federal Circuit case, Ebanks v. Shulkin, which is emphasized by the concurrence in Monk and all but ignored by the plurality. 33 33 See Ebanks v. Shulkin, 877 F.3d 1037 (Fed. Cir. 2017).
A. The Federal Circuit and VA’s low information nature require a more relaxed approach to certification
Generally, there are two reasons that a court relies on when relaxing standards for certain petitioners over others. The first is when stare decisis counsels that one rule is better suited for application to a specific class and not another. The second is when the practical concerns of the parties overwhelm the legal and procedural concerns of the court. In delay claims at the VA, both reasons to lower standards are apparent.
In Ebanks, the Federal Circuit opined that, “the issue [of systemic delay] seems best addressed in the class-action context where the court could consider class-wide relief.” There, the petitioner sought a writ of mandamus after two years of waiting for a hearing to be scheduled before the Board. Despite mooting the issue, the Federal Circuit “question[ed] the appropriateness of granting individual relief to veterans who claim unreasonable delays in VA’s first-come-first-served queue.” Specifically, the Federal Circuit was concerned that addressing the problem on an individual basis would amount to little more than
line-jumping occurs, net gains in the benefits system are lost, and the action ultimately fails to resolve the underlying issues causing the delay. 34 34 See Ebanks v. Shulkin, 877 F.3d 1037, 1039–40 (Fed. Cir. 2017); see also In re Barr Labs., Inc., 930 F.2d 72, 75 (D.C. Cir. 1991). Relief for one veteran through mandamus becomes relief at the expense of another veteran whose claim must be cured through mandamus, which becomes relief at the expense of another veteran, revealing the Federal Circuit’s reasoning for promoting class actions in the context of VA delay.
The Monk plurality also overlooks the practical concerns of veterans and low-information nature of the VA when requiring that veterans affirmatively show the reason for their delay. The veterans’ benefits system has become demonstrably more formal and complex, but despite its growing intricacy and backlog, there have been few developments in the VA’s transparency. 35 35 See James D. Ridgway, The Veterans’ Judicial Review Act Twenty Years Later, 66 N.Y.U. Ann. Surv. Am. L. 251, 253 (2010). Unlike plaintiffs in tort law, where detailed discovery procedures guide the gathering of evidence for litigation, there are few fact-finding methods a veteran can employ when seeking the reasons for his or her individual delay. In fact, when veterans sought procedural protections at the VARO level that mirrored civil litigation, such as
[s]ubpoena power, discovery, pre-decision hearings, and the presence of paid attorneys, the Ninth District felt that
Congress quite plainly intended to preclude any resemblance of civil litigation from VA adjudication. 36 36 Veterans for Common Sense v. Shinseki, 563 F.Supp. 2d 1049, 1064 (N.D. Cal. 2008), affirmed in part, reversed in part on other grounds, 678 F.3d 1013 (9th Cir. 2012). This paradigm eliminates other fact-finding methods, depositions or cross-examinations, which may be necessary to uncover the specific reason for delay.
That leaves both veterans and their representatives very few tools with which to fact-find.One rarely used option is a request under Nohr v. McDonald. 37 37 Nohr v. McDonald, 27 Vet. App. 124 (2014). In Nohr, a veteran, through counsel, requested the submission of
interrogatory type requests to a VHA expert after that expert stated without elaboration,
I recognize my personal limitation . . . when rendering an opinion on the etiology of the veteran’s claimed disability. 38 38 Nohr v. McDonald, 27 Vet. App. 124, 127 (2014). Although the CAVC ultimately held that veterans can engage in fact-finding where they reasonably raise concerns about (1) a VHA expert’s competence, (2) the adequacy of that expert’s opinion, or (3) the VA’s duty to assist, the CAVC chose not to affirm a larger argument that veterans have a Fifth Amendment procedural due process right to confront his or her examiners. In doing so, the holding deferred to the Board to implement statutory and regulatory remedies when Nohr was implicated. This narrow holding left many VA adjudicators reluctant to approve Nohr requests where the examiner does not raise the issue of his or her own limitations. Moreover, this fact-driven inquiry compounds negatively on veterans seeking aggregate resolution for delay because only the third prong relates to a potential reason for delay — VA’s duty to assist. 39 39 See Ellen Brandau, The Future of Interrogatories Under Nohr v. McDonald, 8 Vet. L. Rev. 48 (2016).
A second option for veterans to pursue fact finding is through the Freedom of Information Act (FOIA). Indeed, FOIA requests are more versatile than Nohr requests by having fewer restrictions on the type of documents and departments that the request can target, and they can be utilized by both counsel and lay veterans alike. However, FOIA requests are extremely difficult to use offensively and are more often seen as reactionary. For example, a veteran receives an unfavorable decision from the VARO that relies on an unpublished policy letter. FOIA would be a good way to uncover that policy letter because a specific request can be made. Conversely, when the reason for a delay is unknown, it cannot be targeted by a FOIA request. Public statistics from 2018 show that VA received 24,555 FOIA requests. Of those requests, ~59% were partially denied and another ~23% were denied in full. Only ~17% were granted outright. Even though most of the 24,555 were processed within the calendar year, VA received only 354 appeals of denials. Not only do these number highlight the issue of FOIA’s actual transparency, but its inefficiency sheds light on the administrative burden, only furthering the inaccessibility of FOIA as a tool.
Examples of unreasonable delay that veterans would be unable to unveil range from the trivial to the egregious. In a 2012 report, the Office of the Inspector General (OIG)
found that [VARO’s] did not assign enough staff to process appeals. 40 40 Office of Inspector Gen., Audit of VA Regional Offices’ Appeals Management Processes, U.S. Dep’t Vet. Affairs i, May 30, 2012. More recently, reports on the timeliness of the appeals process showed that the
VBA took an average of 111 to 725 days to complete various phases . . . and . . . found significant periods of inactivity throughout all phases. . . . On average, a single period of inactivity accounted for approximately 45 to 76 percent of the total processing time in each phase. 41 41 Office of Inspector Gen., Review of Timeliness of the Appeals Process, U.S. Dep’t Vet. Affairs ii (Mar. 28, 2018).
For delay claims, the VA’s control over dispositive evidence is juxtaposed by the urgency to help veterans. Instead of being forthright regarding reasons for delay, the VA historically waits until specific policies and practices are published by the OIG.
In addition to the OIG’s findings, litigation proves that many VA personnel have taken the wait times lightly despite the lengthening queue being publicly and closely correlated with rising suicide statistics. Factual findings from 2008, in Veterans for Common Sense v. Peake, show employees attempting coverups in comically hushed tones:
In [one] internal VA email dated February 13, 2008, Dr. Katz wrote:Shh! Our suicide prevention coordinators are identifying about 1,000 suicide attempts per month among the veterans we see in our medical facilities. Is this something we should (carefully) address ourselves in some sort of release before someone stumbles on it?42 42 Veterans for Common Sense v. Peake, 563 F. Supp. 2d 1049, 1063 (N.D. Cal. 2008), aff’d in part, rev’d in part sub nom. Veterans for Common Sense v. Shinseki, 678 F.3d 1013 (9th Cir. 2012).
More findings from VCS prove that many policy initiatives were actively ignored by VA. Key components of the 2004 Mental Health Strategic Plan (MHSP) provide an example, such as
screening veterans at risk, a suicide prevention database, emerging best practices for treatment, and education programs were still at the ‘Pilot Stage’ three years after the MHSP was implemented. 43 43 Veterans for Common Sense v. Peake, 563 F. Supp. 2d 1049, 1065 (N.D. Cal. 2008).
Older litigation from 1987, National Ass’n of Radiation Survivors v. Turnage, highlights the extraordinary lengths VA employees would go to hide their transgressions. 44 44 National Ass’n of Radiation Survivors v. Turnage, 111 F.R.D. 543 (N.D. Cal. 1987). There, two employees in field operations destroyed general files two years or older under the guise of VA
guidelines for document retention. Three more employees
fill[ed] approximately three to four government-issue garbage cans with purged documents. An additional twelve boxes of records were destroyed merely to create additional office space. 45 45 National Ass’n of Radiation Survivors v. Turnage, 111 F.R.D. 543, 546–48 (N.D. Cal. 1987). Importantly, the District Court noted that the timing of many of the destroyed documents coincided with the Plaintiff’s eighth request for production of documents, meaning that the purged records were done so with notice of pending litigation. It follows then that veterans and their representatives are unable to propose or construct classes and subclasses based on a particular policy or practice without complicated discovery to determine the reason for the delay. Because fact-finding is generally a duty to be conducted in the first instance of claims processing, and because of the low-information nature of the VA, it should not be the veteran’s responsibility to discover the reason for delay.
The Monk plurality
silently rejected the Federal Circuit’s opinion in Ebanks after the higher court opined that class-actions would be an efficient way to adjudicate systemic delay. While the strict requirement to show a specific harmful policy resonates throughout Monk, it is important to note that Parsons and the Monk plurality are not binding — their persuasiveness is only a guide for future class certifications. Instead, Godsey shows that the CAVC can conduct limited fact-finding into the reasons for delay in aggregated claims rather than place that burden on the veterans. 46 46 See Monk v. Wilkie, 30 Vet. App. 167, 170–71 (2018); Elkins v. Gober, 229 F.3d 1369, 1377 (Fed. Cir. 2000).
B. Godsey and the Existing Path to Relaxing the Commonality Requirement
Godsey’s resemblance to Monk is nearly ubiquitous. First, Godsey addressed a similar delay issue, alleging that the VA failed to timely certify and transfer appeals to the Board during a process called pre-certification review. Furthermore, Godsey’s class failed to allege a specific policy or practice adversely affecting each member. They also requested similar mandamus relief to remedy the class’s claims. Despite the parallels between the two cases, the majority in Godsey rejected much of Monk’s reasoning, and instead followed the Federal Circuit’s precedent in Ebanks. Essentially holding oppositely to Monk, the Godsey majority did not require veterans to allege a specific policy or practice in order to be bound in commonality under Rule 23(a)(2). The Godsey majority’s solution to the commonality debate, one the dissent adamantly disagreed with, was to sua sponte exercise the power to modify classes under Federal Rule of Civil Procedure 23(c)(5). 47 47 See Fed. R. Civ. P. 23; Godsey v. Wilkie, 31 Vet. App. 207 (2019).
Class modification has a long history of being utilized outside of the veterans context as a tool for equitable relief. Primarily employed by district courts, modifications often look like the court expanding or shrinking the size of the class. Other times the court will create subclasses from the larger class for more manageable litigation. The Supreme Court has historically ruled that the burden is generally on a putative class to construct classes and sub-classes. For example, in United States Parole Commission v. Geraghty, the putative class consisted of federal prisoners challenging the U.S. Parole Commission’s release guidelines. 48 48 United States Parole Comm’n et al. v. Geraghty, 445 U.S. 388, 408 (1980). There, the Supreme Court reasoned that no sua sponte duty existed for a court to modify the class on behalf of the prisoners. However, the rules surrounding prisoners are sufficiently distinguishable from those appropriate for veterans. More applicable comparisons come from more sensitive classes that deal intimately with human health and welfare. As such, comparable class action cases in the Social Security, Medicare, and Medicaid context have shown that courts wield wide latitude when constructing and modifying classes and subclasses. 49 49 See Prado-Steiman v. Bush, 221 F.3d 1266, 1270 (11th Cir. 2000); Richardson v. Perales, 402 U.S. 389, 400–01 (1971).
Although the Supreme Court in Geraghty found no affirmative obligation for courts to construct subclasses, the Eleventh Circuit in Prado-Steiman saw fit to accept that exact burden. In Prado-Steiman, defendants sought review from the Eleventh Circuit over the certification of a single class consisting of all developmentally disabled persons seeking declaratory and injunctive relief from delays and denials of Medicaid benefits. All parties agreed some type of class should have been certified, but the defendants believed the single class was too broad to meet commonality under Rule 23(a)(2). The Eleventh Circuit found that,
[w]hile the alleged injuries of these [claimants] may overlap to some degree, there [were] obvious and important differences that inhibited commonality amongst the class. 50 50 Prado-Steiman v. Bush, 211 F.3d 1266, 1281 (11th Cir. 2000). Recognizing that a compelling public interest permeated the case, the Court proposed breaking the single class into several subclasses to meet commonality requirements. It further remanded the case with instructions for the district court to implement the proposed subclasses. 51 51 See, e.g., Susan J. v. Riley, 254 F.R.D. 439, 451 (M.D. Ala. 2008).
Like the Eleventh Circuit in Prado-Steiman, the CAVC in Godsey also adopted the practice of sua sponte class construction and modification under Rule 23. The initial Godsey class requested that they be certified and bound through the question of whether:
a two-year delay to certify and transfer cases to the Board constitutes a per se violation of class members’ due process rights or is per se unreasonable under TRAC and Martin v. O’Rourke. 52 52 Godsey v. Wilkie, 31 Vet. App. 207, 221 (2019) (citations omitted). The majority noted in Godsey that this proposed class included both claimants in line for pre-certification review as well as those whose claims were subject to
readjudication resulting from pre-certification review. Those veterans among the class whose appeals were at the pre-certification stage for a second time endangered commonality among the rest of the members, and the entire panel agreed that a class comprising of veterans whose claims had already been certified was too broad. Rather than simply deny certification because the petitioners sought too expansive of a class, the majority exercised its discretion to modify the class by
narrow[ing] the class to include only those claimants who have been standing in line waiting more than 18 months since filing their Substantive Appeals. 53 53 Godsey v. Wilkie, 31 Vet. App. 207, 222 n. 3 (2019). Citing a trial court’s general power to amend and adjust proposed classes, the majority specifically found that the CAVC is not bound by a petitioner’s proposed class definition. 54 54 Godsey v. Wilkie, 31 Vet. App. 207, 221 (2019). The choice to modify the class rather than deny certification significantly relaxes Rule 23(a)(2) in the veterans context.
This modified class question glues the members together in a way fit for determination on the merits. However, as the dissent notes, the initial Godsey class should have suffered from the same technical defects of the Monk class. The agency could provide various reasons for delay among the petitioners’ claims, so dissimilarities would eventually poison the commonality of the class. The dissent further emphasized that the class could not be modified and must be denied primarily because petitioners never requested modification from their original proposal. Taking action sua sponte would effectively subvert the commonality analysis that prevented Monk from achieving certification in a similar issue. In this argument, the dissent pertinently identified a
chicken versus egg scenario as the practical consequence of the majority’s decision, reasoning that without a request by the class to modify itself, the court essentially
made a determination on the merits and certified a class based on that determination. 55 55 Godsey v. Wilkie, 31 Vet. App. 207, 231–32 (2019) (Pietsch, J., dissenting). However, the blend of the merits and commonality determinations is weighed too heavily by the dissent, considering that the Supreme Court’s decision in Wal-Mart required courts to conduct a
rigorous analysis that necessarily overlaps the two. The CAVC has treated the commonality requirement, in the context of delay and non-delay claims, differently. Although not a delay claim, in Wolfe v. Wilkie, the CAVC rejected the VA’s argument that when analyzing the commonality of a class, the
answer alone must dispose of the class members claims. The Court reasoned that such standard is too stringent where the putative class does not make an as-applied challenge to the regulation at issue, but instead raises a facial challenge to the regulation’s validity. Thus, the CAVC held that a difference in facts of the Wolfe class does not
stymie certification of this class as they did in the Monk plurality. 56 56 See Wolfe v. Wilkie, 32 Vet. App. 1, 28–29 (2019).
Even though the Supreme Court in Geraghty required a putative class to propose and construct subclasses for prisoners, veterans more closely resemble the Medicaid claimants in Prado-Steiman where the Eleventh Circuit’s rigorous analysis led it sua sponte modify and create subclasses for sensitive claimants. And while Medicaid applicants in Prado-Steiman traverse an adversarial district court, veterans are embroiled in a uniquely claimant-friendly adjudicatory system. Veterans are thus distinguishable from both prisoners in Geraghty and even from Medicaid claimants in Prado-Steiman and should command a lesser standard for finding commonality among a class than prescribed by Monk. The Godsey Court’s insistence that the CAVC is not bound by a class’ proposed definition, coupled with its willingness to extend the principle of class modification to meet commonality requirements, effectively lowers that standard.
IV. Class Modification in Practice
Although not a delay claim, in Skaar v. Wilkie, the CAVC has already illustrated its willingness to extend the principle of class modification. In Skaar, veterans who participated in the nuclear cleanup of Palomares, Spain collectively challenged the VA’s reliance on the Air Force’s dose methodology when calculating exposure to radiation in compliance with 38 C.F.R. § 3.311. 57 57 Skaar v. Wilkie, 32 Vet. App. 156, 168 (2019). Specifically, each class member claimed that they were harmed by VA’s failure to rely on
sound scientific and medical evidence.
The question of whether VA’s use of a certain methodology is prejudicial on its face strikes at a dispositive legal issue central to each member’s claim. However, in briefing, the Secretary argued that a dissimilarities analysis highlighted impediments to the proposed class definition which would ultimately erode commonality. Indeed, the petitioner’s initial proposal was expansive and not atypical of an attempt to incorporate as many class members in the action as possible, including:
veterans whose claims for . . . benefits related to exposure to ionizing radiation at Palomares the VA has denied at any level . . . except for those who have appealed to [the CAVC] and received a decision for which the mandate has issued;
veterans whose claims the [VARO] or [Board] has denied and for which the deadline for appeal has expired, as well as veterans whose claims are currently pending before a [decision review officer] or the [Board] after an initial [VARO] denial;and
Palomares veterans with an appeal currently pending before [the CAVC].58 58 Skaar v. Wilkie, 32 Vet. App. 156, 171–72 (2019).
The proposed class also includes “veterans with claims that have not yet been filed at the RO,” including
those who have not filed a claim for an existing condition, including because they are aware of the VA’s history of denial of Palomares veterans’ claims or the methodology used to calculate dose exposure” and “those who have only recently developed a radiogenic condition, and those whose claims have been delayed at the RO. 59 59 Skaar v. Wilkie, 32 Vet. App. 156, 171–72 (2019).
Eventually, these proposed classes were modified to include
[a]ll U.S. veterans who were present [the Palomares Cleanup], and whose application for service-connected disability compensation based on exposure to ionizing radiation the VA has denied or will deny by relying, at least in part, on the findings of dose estimates requested under 38 C.F.R. § 3.311. 60 60 Skaar v. Wilkie, 32 Vet. App. 156, 189 (2019).
Notably, the CAVC excluded those past or expired claimants whose claims were denied before reaching the Board but did not perfect an appeal of that denial. The CAVC also excluded those who did have final Board decisions but failed to file a Notice of Appeal to the CAVC.
The majority in Skaar recognized that it could not exercise jurisdiction over those abandoned claims. In removing those claims from its purview, the CAVC was very careful to note that these modified classes are not formal subclasses, but are “subgroups merely for purposes of analyzing our jurisdiction as to each [putative member] . . .” As in Godsey, where veterans at the pre-certification stage for a second time endangered the possibility of a common resolution, the Skaar Court sua sponte modified the class to exclude abandoned claims so as to provide a path to resolution for the rest of the claimants. In doing so, the CAVC clearly engaged in a form of first instance fact finding, much like what occurs at the agency levels.
Where there are few tools at the disposal of veterans and their counsel to fact find in a timely manner, Skaar and Godsey show that the CAVC is willing to conduct necessary discovery on behalf of veterans. These holdings are rather novel because, typically, the CAVC is barred from considering facts that were not before the Board prior to an appeal to the Court. 61 61 38 U.S.C. § 7252(a). It is from this principle that the Monk plurality required veterans to affirmatively plead the reasons for delay before the CAVC could hear them. However, Skaar shows that the practical effect of certifying class actions at an appellate court is the need for the court to employ self-imposed fact-finding. While Godsey introduced the idea that classes should be modified where dissimilarities are the only factors impeding commonality, the Skaar Court further extended the Court’s fact-finding powers, essentially wielding them as a tool to self-police its own jurisdiction during the certification process.
At this juncture, the waters of class action litigation as applied to veterans are a little less murky. The decision to adopt rules for class actions consistent with the FRCP reduces uncertainty in the initial stages of certification, and it gives claimants freedom to assume that the rule does not depart in meaning from the Federal Rules. However, the initial application of Rule 23(a)(2) was too strict. The affirmative duty that Geraghty, Parsons, the Monk plurality, and the Godsey dissent impute onto veterans to construct a class based on a specific policy is unduly burdensome. This paradigm may be supported by precedent but certainly not the practical needs of litigants seeking class certification at an appellate court. It neglects the low-information nature of the VA and assumes that veterans can uncover the reason for their delay. Moreover, many of these class scenarios involve sufficiently distinguishable populations from veterans.
Instead, the history of class modification provides a naturally existing framework to relaxing Rule 23(a)(2). Like the Ninth Circuit in Prado-Steiman, the modification doctrine specifically counseled the CAVC in Godsey to conduct the rigorous analysis required by Wal-Mart in a way that illuminates a common class even where none is proposed. In Skaar, the CAVC has shown that it plans to continue that practice. Not only are human health and welfare concerned, but a lack of transparency within the adjudications process highlights the inability to uncover the reasons for their individual case’s delay. This information gap obligates the CAVC to adopt an agency-like power in the first instances of fact-finding for classes alleging delay and sua sponte modify classes where justice so requires.