NATIONAL ORGANIZATION OF VETERANS' ADVOCATES, INC., Peter Cianchetta, Michael Regis, Andrew Tangen, Petitioners v. SECRETARY OF VETERANS AFFAIRS, Respondent
At the heart of the government's scheme for awarding disability benefits to veterans is a rating schedule. The Department of Veterans Affairs adopted this rating schedule to standardize the evaluation of how severely diseases and injuries resulting from military service impair veterans' earning capacity. 38 C.F.R. § 4.1. The rating schedule is, in turn, divided into diagnostic codes that provide disability ratings for various symptoms or conditions.
National Organization of Veterans' Advocates, Inc., Peter Cianchetta, Michael Regis, and Andrew Tangen petition this court under 38 U.S.C. § 502 to review the VA's interpretation of two of these diagnostic codes: DCs 5055 and 5257, both found at 38 C.F.R. § 4.71a. The VA set out its interpretation of DC 5055 in Agency Interpretation of Prosthetic Replacement of a Joint, 80 Fed. Reg. 42,040 (July 16, 2015) (the “Knee Replacement Guidance” or “Guidance”), and VA Adjudication Procedures Manual M21-1 Section III.iv.4.A.6.a (the “Knee Replacement Manual Provision”). The VA set out its interpretation of DC 5257 in Manual Section III.iv.4.A.6.d (the “Knee Joint Stability Manual Provision”).
For the reasons provided below, we conclude that the Knee Replacement Manual Provision is not a reviewable agency action. We also hold that the Knee Replacement Guidance is arbitrary and capricious under the controlling precedent of Hudgens v. McDonald, 823 F.3d 630 (Fed. Cir. 2016). Finally, we dismiss the challenge to the Knee Joint Stability Manual Provision as moot. Accordingly, we grant-in-part and dismiss-in-part the petition.
I. The VA's Interpretation of DC 5055
Petitioners seek review of two interpretive rules. To introduce the VA's interpretation of DC 5055, we must turn back to the claim at issue in Hudgens. In that case, the VA regional office (“RO”) denied Michael A. Hudgens, a U.S. Army veteran, a 100-percent disability evaluation for his partial prosthetic knee replacement under DC 5055 because the RO found that DC 5055 applied only to total knee replacements. Hudgens, 823 F.3d at 632–33. The Board of Veterans' Appeals and the United States Court of Appeals for Veterans Claims both affirmed the RO's conclusion that DC 5055 did not apply to Mr. Hudgens's partial knee replacement claim. Id. at 633–34. Mr. Hudgens then appealed to this court. Id. at 634.
On July 16, 2015, twelve days before the Secretary's final brief in Hudgens was due with this court, the VA published the Knee Replacement Guidance. Id. The Guidance stated that the VA was providing notice of the agency's “longstanding interpretation of DCs 5051 to 5056” as providing for a 100-percent evaluation “when the total joint, rather than the partial joint, has been replaced by a prosthetic implant.” 80 Fed. Reg. at 42,040. The VA also announced in the Guidance that an “explanatory note” would be added to 38 C.F.R. § 4.71a stating that the “term ‘prosthetic replacement’ in diagnostic codes 5051 through 5056 means a total replacement of the named joint.”1 Id. at 42,041.
In Hudgens, we nevertheless reversed the judgment of the Veterans Court and remanded for further proceedings. 823 F.3d at 640. We held that DC 5055 “does not unambiguously exclude [partial knee] replacements.” Id. at 637 (emphasis omitted). We further concluded that the Secretary's interpretation of DC 5055 could not be afforded Auer deference for two reasons. First, the Secretary's interpretation “conflict[ed]” with “numerous inconsistent rulings by the Board” holding that partial knee replacements could be evaluated under DC 5055. Id. at 638–39. Second, the Knee Replacement Guidance was a “post hoc rationalization” “conveniently adopted to support the Veterans Court's interpretation in this case.” Id. at 639. Finally, we held that Mr. Hudgens's “interpretation of DC 5055 is permitted by the text of the regulation,” meaning that we had to apply the pro-veteran canon, see id.; see also Brown v. Gardner, 513 U.S. 115, 117–18, 115 S.Ct. 552, 130 L.Ed.2d 462 (1994), and “resolve any doubt in the interpretation of DC 5055 in his favor,” Hudgens, 823 F.3d at 639. His claim, therefore, could be evaluated under DC 5055. Id.
On November 21, 2016, six months after our decision in Hudgens, the VA informed RO adjudicators of how the agency intended to reconcile our decision in that case with the Knee Replacement Guidance. J.A. 4, 28. In the Knee Replacement Manual Provision, the VA directed RO adjudicators to not evaluate under DC 5055 any claims for partial knee replacements “filed and decided on or after July 16, 2015.” J.A. 28. Claims filed before July 16, 2015, and pending as of that date were to be evaluated under DC 5055 if doing so would be more favorable than evaluating the same claims under another applicable diagnostic code. Id. Finally, claims filed before July 16, 2015, and adjudicated before that date were not to be revised. Id.
Four years later, the VA amended DC 5055 following notice-and-comment to “clarify VA's intent to provide a minimum evaluation following only total joint replacement.” Schedule for Rating Disabilities: Musculoskeletal System and Muscle Injuries, 85 Fed. Reg. 76,453, 76,454, 76,456 (Nov. 30, 2020). The change was effective February 7, 2021. Id. at 76,453. On February 8, the VA rescinded the Knee Replacement Manual Provision. J.A. 246–47. But “the new Manual provisions still reference the historical Knee Replacement Manual Provision because adjudicators use the historical guidance to rate claims that were pending as of February 7, 2021[,] and that include rating periods prior to that date.” Resp't's Br. 12; see J.A. 318.
II. The VA's Interpretation of DC 5257
The second rule was stated in the Knee Joint Stability Manual Provision. When this Manual provision was promulgated in 2018, J.A. 111, 113, DC 5257 assigned a 10-percent rating for “Slight” knee instability, a 20-percent rating for “Moderate” instability, and a 30-percent rating for “Severe” instability. 38 C.F.R. § 4.71a (2018). The Knee Joint Stability Manual Provision directed RO adjudicators to determine whether a claimant had slight, moderate, or severe instability by measuring the amount of movement in the joint. J.A. 113. The Manual provision associated slight instability with 0–5 millimeters of joint translation, moderate instability with 5–10 millimeters of joint translation, and severe instability with 10–15 millimeters of joint translation. Id. On January 21, 2021, following our decision in National Organization of Veterans' Advocates, Inc. v. Secretary of Veterans Affairs, 981 F.3d 1360 (Fed. Cir. 2020) (en banc) (“NOVA 2020”), the VA rescinded the Knee Joint Stability Manual Provision. J.A. 183, 228.
III. The Instant Appeal
On January 3, 2020, NOVA filed a petition for review of the Knee Replacement Manual Provision and the Knee Joint Stability Manual Provision. NOVA 2020, 981 F.3d at 1365–66. NOVA later amended the petition to add Messrs. Cianchetta, Regis, and Tangen—three members of NOVA—as petitioners and to challenge the Knee Replacement Guidance. Id. at 1368. In NOVA 2020, we held that NOVA has associational standing to challenge the Guidance and both Manual provisions. Id. at 1371. We further held that we have jurisdiction under 38 U.S.C. § 502 to review the VA's interpretation of DCs 5055 and 5257.2 Id. at 1378, 1382. It was left for a merits panel to determine whether the Knee Replacement Guidance or the Knee Replacement Manual Provision constitutes an independently reviewable interpretive rule and to render a decision on the merits of the petition. Id. at 1383, 1386.
I. The Knee Replacement Manual Provision
We begin by resolving the jurisdictional question left open by NOVA 2020. We hold that the Knee Replacement Guidance—not the Knee Replacement Manual Provision—constitutes the final agency action subject to review under § 502. As we explained in NOVA 2020: “Manual provisions that merely republish prior agency interpretations or restate existing law ․ are not reviewable under section 502.” Id. at 1382.
The Knee Replacement Guidance predates the Knee Replacement Manual Provision, and the Manual provision makes no “substantive change” to the Guidance. Id.; see J.A. 1, 25, 28–29. The Manual provision takes as its key date July 16, 2015, the day the VA promulgated the Guidance. J.A. 1, 28. For partial knee replacement claims filed on or after that date, the Manual provision directs RO adjudicators not to award evaluations under DC 5055, thereby implementing the rule put forward in the Guidance. J.A. 1, 28. The Manual provision even explains that its rule stems from the explanatory note added by the Guidance. J.A. 28. Under the Manual provision, partial knee replacement claims filed before July 16, 2015, may still be evaluated under DC 5055, such that the Knee Replacement Guidance has no effect on their adjudication. J.A. 28–29.
Petitioners contend that this “temporal limitation is not contained in the Knee Replacement Guidance,” so the Manual provision is substantively different from the Guidance. Pet'rs' Br. 50. But there is no suggestion in the Guidance that it is to have retroactive effect; indeed, the Guidance states that it has an effective date of July 16, 2015. J.A. 1. Thus, the Manual provision does not provide instructions to RO adjudicators that are substantively inconsistent with the Knee Replacement Guidance. Because it “merely republish[es]” the Guidance, the Knee Replacement Manual Provision is not a final agency action subject to review under § 502. NOVA 2020, 981 F.3d at 1382–83.
II. The Knee Replacement Guidance
Having concluded that the Knee Replacement Guidance is the reviewable agency action, we now consider whether the Guidance is arbitrary and capricious. We conclude that it is.
A. Standard of Review
“We review petitions under 38 U.S.C. § 502 in accordance with the standard set forth in the Administrative Procedure Act (‘APA’), 5 U.S.C. §§ 701–706.” Paralyzed Veterans of Am. v. Sec'y of Veterans Affs., 345 F.3d 1334, 1339 (Fed. Cir. 2003) (citing Nyeholt v. Sec'y of Veterans Affs., 298 F.3d 1350, 1355 (Fed. Cir. 2002)). Under the APA, we “hold unlawful and set aside agency action” that is “arbitrary [and] capricious,” is “not in accordance with law,” or is promulgated “without observance of procedure required by law.” 5 U.S.C. §§ 706(2)(A), (D).
B. The Text of DC 5055
We are asked to determine whether the VA's interpretation of DC 5055 is arbitrary and capricious or contrary to law. As always, we begin our analysis by looking to the regulatory text. See Chase Bank USA, N.A. v. McCoy, 562 U.S. 195, 204, 131 S.Ct. 871, 178 L.Ed.2d 716 (2011). “[I]f there is only one reasonable construction of a regulation,” then a court should not defer to any conflicting agency interpretation. Kisor v. Wilkie, ––– U.S. ––––, 139 S. Ct. 2400, 2415, 204 L.Ed.2d 841 (2019).
We look to the text of DC 5055 on the date that the VA promulgated the Knee Replacement Guidance. See Chase Bank, 562 U.S. at 204, 131 S.Ct. 871. On July 16, 2015, DC 5055 stated:
5055 Knee replacement (prosthesis).
Prosthetic replacement of knee joint:
For 1 year following implantation of prosthesis ․ 100
With chronic residuals consisting of severe painful motion or weakness in the affected extremity ․ 60
With intermediate degrees of residual weakness, pain or limitation of motion rate by analogy to diagnostic codes 5256, 5261, or 5262.
Minimum rating ․ 30
38 C.F.R. § 4.71a (2015).
In Hudgens, we held that DC 5055 does not unambiguously exclude partial knee replacements. 823 F.3d at 637. Accordingly, the text of DC 5055 does not resolve whether the Knee Replacement Guidance is arbitrary and capricious. Id. at 637–38. The Secretary argues that we are not bound by this holding in Hudgens because DC 5055 has subsequently been “clarified” by the addition of the explanatory note. Resp't's Br. 31. The Secretary asserts that the explanatory note is “plain and unambiguous” in stating that “the term ‘prosthetic replacement’ in diagnostic codes 5051 through 5056 means a total replacement of the named joint.” Id. at 30 (quoting 38 C.F.R. § 4.71a, DCs 5051–56, Note (2020)).
We reject this circular argument. We are evaluating whether the Guidance constitutes a valid interpretation of DC 5055. The Guidance itself inserted the explanatory note into DC 5055. J.A. 1–3. The Secretary would have us hold that the Guidance articulates the only reasonable reading of DC 5055 because the Guidance itself says so. See generally Resp't's Br. 29–31. That cannot be correct. Indeed, the Secretary's argument contravenes a basic tenet of administrative law. Agencies must “use the same procedures when they amend or repeal a rule as they used to issue the rule in the first instance.” Perez v. Mortg. Bankers Ass'n, 575 U.S. 92, 101, 135 S.Ct. 1199, 191 L.Ed.2d 186 (2015); see 5 U.S.C. § 551(5). The VA promulgated DC 5055 following notice-and-comment rulemaking procedures in 1978. Updating the Schedule for Rating Disabilities, 43 Fed. Reg. 45,348, 45,348–50 (Oct. 2, 1978). Therefore, the Secretary cannot have amended DC 5055 without going through notice-and-comment.
In Hudgens, we considered the same version of DC 5055 that existed when the VA promulgated the Knee Replacement Guidance. Compare 38 C.F.R. § 4.71a (2015), with Hudgens, 823 F.3d at 632 (quoting 38 C.F.R. § 4.71a). We thus follow Hudgens and conclude that DC 5055 is ambiguous as to whether it includes partial knee replacements.
We next consider, as we did in Hudgens, whether we must defer to the Secretary's interpretation under Auer v. Robbins, 519 U.S. 452, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997). We conclude that Hudgens is still controlling precedent and so we cannot afford the Secretary's interpretation Auer deference.
In Hudgens, we gave two reasons for why the Secretary's interpretation of DC 5055—excluding partial knee replacements—was not entitled to Auer deference. 823 F.3d at 638–39. First, “the agency's interpretation conflict[ed] with a prior [agency] interpretation”—namely, “numerous inconsistent rulings by the Board.” Id. at 638 (second brackets in original) (quoting Christopher v. SmithKline Beecham Corp., 567 U.S. 142, 155, 132 S.Ct. 2156, 183 L.Ed.2d 153 (2012)). Second, the Secretary's interpretation was a “post hoc rationalization” “adopted to support the Veterans Court's interpretation.” Id. at 639. The second reason regarding post hoc rationalization is not relevant here as we are addressing only the interpretation put forward in the Knee Replacement Guidance, not a previous agency interpretation that the Secretary is justifying with the Knee Replacement Guidance. See Oral Arg. at 01:47–02:07, available at https://oralarguments.cafc.uscourts.gov/default.aspx?fl=20-1321_12092021.mp3 (“Q: [G]oing forward, if [the Guidance] therefore is not to advance a defense of past agency action, we are now talking about using it to defend subsequent agency action. To me, that's a real distinction that has heft, right? A [from petitioners]: I agree it's a distinction, your honor.”). Therefore, this leaves our first reason in Hudgens that “numerous inconsistent rulings by the Board” foreclose the VA's interpretation. 823 F.3d at 638.
The Secretary contends that we should not characterize Board interpretations “as representing the agency's official position.” Resp't's Br. 33–35. The Secretary notes that “to receive Auer deference, ‘the interpretation must at least emanate from those actors, using those vehicles, understood to make authoritative policy in the relevant context.’ ” Id. at 34 (internal brackets omitted) (quoting Kisor, 139 S. Ct. at 2416). The Board is not such an authoritative actor, the Secretary asserts, because it issues more than 100,000 non-precedential decisions a year where the judges act individually rather than in panels. Id. at 35 (citing Board of Veterans' Appeals Annual Report to Congress (FY 2020), https://www.bva.va.gov/docs/Chairmans_Annual_Rpts/BVA2020AR.pdf). In the Secretary's view, these facts, together with our statement in NOVA 2020 that Board decisions “appear not to be entitled to Auer deference,” 981 F.3d at 1382 n.14, “erodes the foundation” of our holding in Hudgens, Oral Arg. at 29:50–52.
The Secretary's argument is beside the point. Neither in Hudgens nor in this case is the issue whether Board decisions are entitled to Auer deference. See id. at 17:28–18:00 (petitioners emphasizing that they are not arguing that Board decisions should be entitled to Auer deference). Rather, the issue is whether the existence of conflicting prior Board decisions can preclude the application of Auer deference to the subsequent VA interpretation in the Knee Replacement Guidance. As we explain below, no relevant law or facts have changed since our decision in Hudgens, and so we continue to be bound by our conclusion that prior conflicting Board decisions interpreting DC 5055 preclude application of Auer deference to the Knee Replacement Guidance.
Since our decision in Hudgens, the Supreme Court has addressed Auer deference in Kisor. In Kisor, the Court found “it worth reinforcing some of the limits inherent in the Auer doctrine.” 139 S. Ct. at 2415. First, the Supreme Court held that before applying Auer deference, courts must “carefully consider the text, structure, history, and purpose of a regulation” and conclude that “the regulation is genuinely ambiguous.” Id. Second, the agency's interpretation must also be “reasonable” and “come within the zone of ambiguity the court has identified after employing all its interpretive tools.” Id. at 2415–16. Third, “a court must make an independent inquiry into whether the character and context of the agency interpretation entitles it to controlling weight.” Id. at 2416. This third part of the Auer inquiry is relevant here. The Supreme Court gave a list of several “especially important markers for identifying when Auer deference is and is not appropriate” given the character and context of the interpretation. Id.
The first of these markers, the Court explained, is that only “authoritative” agency interpretations should be afforded Auer deference. Id. “That constraint follows from the logic of Auer deference—because Congress has delegated rulemaking power, and all that typically goes with it, to the agency alone.” Id. In other words, Auer deference is available only when agencies exercise their “delegated rulemaking power.”
But not all agency actions exercising such power can receive Auer deference. See Kristin E. Hickman & Richard J. Pierce, Jr., Administrative Law § 3.8.3 (6th ed. 2022) (summarizing that under Kisor an agency interpretation being “authoritative” is one of five conditions that must be met before that interpretation can receive Auer deference). The Supreme Court proceeded to separately summarize case law holding that an agency interpretation should not receive Auer deference when it conflicts with a prior interpretation and “creates ‘unfair surprise’ to regulated parties.” Kisor, 139 S. Ct. at 2417–18 (quoting Long Island Care at Home, Ltd. v. Coke, 551 U.S. 158, 170, 127 S.Ct. 2339, 168 L.Ed.2d 54 (2007)). Such agency actions do not reflect the “fair and considered judgment” of the agency. Id. at 2417 (quoting Christopher, 567 U.S. at 155, 132 S.Ct. 2156). Where there is such an “upending of reliance,” the lack of “ ‘fair warning’ outweigh[s] the reasons to apply Auer.” Id. at 2418 (quoting Christopher, 567 U.S. at 156, 132 S.Ct. 2156). Accordingly, the primary concern of this constraint on Auer deference is the expectations that the agency has previously engendered.
Indeed, an agency can create such expectations even if it has not previously put forward an “authoritative” agency interpretation. Id. at 2418 (“[T]he upending of reliance may happen without such an explicit interpretive change.”); see also Romero v. Barr, 937 F.3d 282, 291 (4th Cir. 2019) (“[T]he upending of reliance may happen without such an explicit interpretive change. Rather, an agency may—instead of issuing a new interpretation that conflicts with an older one—set forth an interpretation for the first time that is contrary to an established practice to which the agency has never objected.” (internal citation omitted)). As Kisor explains, in Christopher, the Supreme Court “refused to defer to an interpretation that would have imposed retroactive liability on parties for longstanding conduct that the agency had never before addressed.” Kisor, 139 S. Ct. at 2418 (citing Christopher, 567 U.S. at 155–56, 132 S.Ct. 2156). In other words, even the absence of prior agency action can cause a new interpretation to be an “upending of reliance,” preventing that interpretation from receiving Auer deference. Christopher, 567 U.S. at 155–58, 132 S.Ct. 2156. Kisor merely reiterated that conflicting agency interpretations cannot receive Auer deference because they cause “unfair surprise” or an “upending of reliance.” 139 S. Ct. at 2417–18 (quoting Long Island, 551 U.S. at 170, 127 S.Ct. 2339); accord Charles H. Koch, Jr. & Richard Murphy, 3 Administrative Law & Practice § 10:26 (3d ed. 2022) (“In Christopher ․ the circumstance that mattered most to the Court was the element of unfair surprise.”). Kisor did not change the exception to Auer deference applied in Hudgens. See Hudgens, 823 F.3d at 638–39 (citing Christopher, 567 U.S. at 155, 132 S.Ct. 2156).
The fact that the Board, as it did before Hudgens, issues numerous non-precedential decisions by single judges each year does not alter our decision. Of the relevant Board decisions issued before the VA promulgated the Guidance, the vast majority applied DC 5055 to partial knee replacements. Id. at 637–38 & n.3 (noting Mr. Hudgens's argument that 17 out of 21 Board decisions issued prior to Hudgens interpreted DC 5055 to cover partial knee replacements); Hudgens v. Gibson, 26 Vet. App. 558, 566 (2014) (“Gibson”) (Kasold, C.J., dissenting) (noting that there were “at least 11 Board decisions that have interpreted DC 5055 to cover partial knee replacements and only 3 that have interpreted it as limited solely to total knee replacements”), rev'd sub nom. Hudgens v. McDonald, 823 F.3d 630 (Fed. Cir. 2016). And although Board decisions are non-precedential and issued by single judges, they are—as we held in Hudgens—“the final decision[s] for the Secretary on all questions in matters affecting the provision of benefits” and provide persuasive authority to the Veterans Court on the interpretation of regulations. Hudgens, 823 F.3d at 638 (quoting Gibson, 26 Vet. App. at 566 (Kasold, C.J., dissenting) (citing 38 U.S.C. § 7104(c))). As such, there is “weight accorded to Board interpretations of VA regulations” that can engender reliance interests and foreclose application of Auer deference to the later conflicting Knee Replacement Guidance.3 Id.
We are thus still bound by the conclusion in Hudgens that the interpretive rule embodied in the Knee Replacement Guidance cannot qualify for Auer deference because it conflicts with most prior Board decisions interpreting DC 5055. See id. at 638–39.
D. Pro-Veteran Canon of Construction
We also must follow Hudgens, apply the pro-veteran canon, and defer to Petitioners' interpretation of DC 5055.4 We held in Hudgens that “[e]ven if the government's asserted interpretation of DC 5055 is plausible, it would be appropriate under [the pro-veteran canon] only if the [regulatory] language unambiguously supported the government's interpretation.” Id. at 639 (cleaned up) (quoting Sursely v. Peake, 551 F.3d 1351, 1357 (Fed. Cir. 2009)). Petitioners' interpretation of DC 5055, which is the same as that of Mr. Hudgens, “is permitted by the text of the regulation.” Id. “DC 5055 is under the heading ‘Prosthetic Implants’ and merely lists a schedule of ratings for the condition ‘Knee replacement (prosthesis),’ without elaboration or limitation of the condition.” Id. (citing 38 C.F.R. § 4.71a (2015)). The Secretary argues that we are not bound by the analysis of the pro-veteran canon in Hudgens only because the VA has subsequently added the explanatory note. Resp't's Br. 36–37. But we have already concluded that the explanatory note cannot influence our interpretation of DC 5055. So, we follow Hudgens, apply the pro-veteran canon, and resolve any doubt in Petitioners' favor.
* * *
For these reasons, we conclude that the Secretary's interpretation of DC 5055 is arbitrary and capricious and vacate the Knee Replacement Guidance.
III. The Knee Joint Stability Manual Provision
Finally, we turn to the Knee Joint Stability Manual Provision. The parties agree that this issue is moot because the Secretary rescinded the Manual provision. Pet'rs' Br. 54–55; Resp't's Br. 55–56. We concur. Although Petitioners ask us to declare that the Manual provision is invalid ab initio, see Pet'rs' Br. 55, we do not have the authority to make such a ruling on the merits when the issue is moot, see Powell v. McCormack, 395 U.S. 486, 496 n.7, 89 S.Ct. 1944, 23 L.Ed.2d 491 (1969).
Accordingly, we grant the petition as to the Knee Replacement Guidance and dismiss the petition as to the Knee Replacement Manual Provision and Knee Joint Stability Manual Provision.
At issue here is the Secretary's guidance interpreting diagnostic code 5055 (“DC 5055”) as limited to total knee replacements. See 80 Fed. Reg. 42040 (July 16, 2015) (“Guidance”). The Secretary published the Guidance during a prior appeal, Hudgens v. McDonald, 823 F.3d 630 (Fed. Cir. 2016). There, when urged to apply the Guidance retrospectively, we withheld deference under Auer v. Robbins, 519 U.S. 452, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997), because we concluded that the Guidance was a post-hoc rationalization in conflict with prior Board decisions. Hudgens, 823 F.3d at 638–39. Then, with Auer off the table, we resorted to the pro-veteran canon, which instructs that “interpretive doubt is to be resolved in the veteran's favor,” Brown v. Gardner, 513 U.S. 115, 118, 115 S.Ct. 552, 130 L.Ed.2d 462 (1994). Invoking that rule, Hudgens held for the veteran solely because his interpretation was “permitted by the text of the regulation.” 823 F.3d at 639 (emphasis added). The majority concludes it is compelled to follow the same exact path laid out in Hudgens. I disagree. And, as I read it, so did Hudgens—which explicitly suggested that going forward the Secretary was free to do what he did here. Even if the majority is correct that DC 5055 is ambiguous, which I take as given for purposes of this dissent, our two reasons for avoiding Auer in Hudgens do not apply here, where we instead confront the Guidance's prospective application. And the pro-veteran canon does not preclude deference here either. I respectfully dissent.
I begin with Hudgens and my disagreement that it resolves this case. In Hudgens, we found DC 5055 ambiguous. Id. at 637. We withheld Auer deference from the Secretary's interpretation, however, which he published during that appeal. Id. at 638. We did so for two reasons, both drawn from the Supreme Court's decision in Christopher v. SmithKline Beecham Corp., 567 U.S. 142, 155, 132 S.Ct. 2156, 183 L.Ed.2d 153 (2012). Because I conclude that neither of those reasons applies in the context of this case, I disagree that Hudgens precludes us from deferring to the Guidance, and I ultimately conclude that's what we should do.
One reason we gave for declining to defer in Hudgens was that the Guidance was an improper post-hoc rationalization. 823 F.3d at 639. The majority acknowledges, and I agree, that this rationale doesn't apply in the circumstances of this case. Maj. ––––. So that pillar of Hudgens is gone.
That “leaves our [other] reason” for withholding deference in Hudgens, Maj. ––––—i.e., that the Guidance “conflicts with a prior [agency] interpretation.” 823 F.3d at 639 (alteration in original) (quoting Christopher, 567 U.S. at 155, 132 S.Ct. 2156). More than six years ago, in Hudgens, we looked backwards and noted that prior Board decisions “favor[ed] Mr. Hudgens's view.” Id. “The precise number,” we continued, was “not important.” Id. at 638 n.3. Perhaps it was “11 out of 14,” id. at 637, or even “17 out of 21 according to [Mr. Hudgens's] count,” id. at 638 n.3. “What [was] important,” we said, was that these embodied “the vast majority of Board decisions considering the question.” Id.
The “primary concern” underlying this rationale was “the expectations that the agency ha[d] previously engendered.” Maj. ––––. That is, “unfair surprise,” upending of “reliance,” and lack of “fair warning” ultimately outweighed the reasons to apply Auer. Christopher, 567 U.S. at 156–57, 132 S.Ct. 2156; see Maj. ––––. This remaining pillar of Hudgens, the majority believes, still stands. In my view, it doesn't. Specifically, this Hudgens rationale is inapplicable for a straightforward reason: in Hudgens, the Secretary tried to apply the Guidance retrospectively to Mr. Hudgens's claim, but it now applies only prospectively—signifying both the presence of “fair warning” and the absence of “unfair surprise” or upended “reliance” here. As the majority recognizes, the Secretary implemented Hudgens by indicating that DC 5055 would cover both partial and total knee replacements for claims filed before the Guidance (like Mr. Hudgens's) but would cover only total knee replacements (per the Guidance) for future claims (like the relevant claims here). Maj. ––––. So, where's the unfair surprise? The upending of reliance? The lack of fair warning? There is none.
The majority downplays this distinction and instead relies on the same handful of “[d]ecisions issued before the [Secretary] promulgated the Guidance,”1 Maj. –––– (emphasis added), which petitioners say “still create the conflict today,” Oral Arg. at 10:35–40, No. 20-1321.2 I doubt the majority would say that, whenever the Secretary corrects a Board interpretation, Auer is forever unavailable; indeed, that position seems contrary to Hudgens itself: “If the Secretary is dissatisfied with the Board's interpretation ․, the Secretary may instruct the Board regarding what the Secretary believes is the correct interpretation.” Hudgens, 823 F.3d at 638.
My conclusion above—that neither of the two Auer-defeating rationales of Hudgens applies—leaves me in quite a different place from the majority. Because the majority calls upon the pro-veteran canon after taking Auer off the table, Maj. –––– – ––––, it does not find itself caught between two rules that “counsel contrary outcomes.” See Hudgens, 823 F.3d at 639 n.5. It erroneously avoids that predicament the same way Hudgens did, by deciding there are “firm grounds upon which to conclude that Auer deference does not apply.” Id.
Having concluded that those grounds (whatever their initial firmness) have given way, however, I do not have the same luxury. I must answer the “difficult and unresolved” question of which of these two ambiguity-resolving rules “gets triggered first”—Auer or the pro-veteran canon. See Kisor v. McDonough, 995 F.3d 1347, 1358 (Fed. Cir. 2021) (Prost, C.J., concurring in denial of rehearing en banc) (“Prost Kisor Concurrence”). At least regarding the ambiguity presented in this case, I conclude the answer is Auer.
I begin with two preliminary points. One, the Supreme Court has never applied the pro-veteran canon to a regulation, and Hudgens was the only decision of this court, until today, to do so. As I've said before, “if the pro-veteran canon is based on the theory that it is a proxy for congressional intent, one wonders why it should apply to regulations as well as statutes, or at least whether it would apply with equal force.” Prost Kisor Concurrence, 995 F.3d at 1349 n.2. Neither Hudgens nor the majority attempts to justify why it should so apply.
Two, even if I agreed with the majority that Auer is off the table and the canon is the appropriate tool to employ here, the majority's application of the canon—hinging on Hudgens's statement that the veteran's interpretation is “permitted by the text of the regulation,” is flawed. Maj. –––– (emphasis added) (quoting 823 F.3d at 639). Hudgens, importantly, was decided before the Supreme Court decided Kisor v. Wilkie. On remand following the Court's decision, we said the pro-veteran canon “does not apply unless ‘interpretive doubt’ is present,” and that “[t]hat precondition is not satisfied where a sole reasonable meaning is identified through the use of ordinary textual analysis tools[ ] before consideration of the pro-veteran canon.” Kisor v. McDonough, 995 F.3d 1316, 1325–26 (Fed. Cir. 2021) (quoting Brown, 513 U.S. at 117–18, 115 S.Ct. 552), cert. denied, ––– U.S. ––––, 142 S. Ct. 756, 211 L.Ed.2d 474 (2022).3 At the same time, we rejected the view that applying the canon is proper so long as the text does not “preclude[ ]” or “expressly exclude the veteran's interpretation.” See id. at 1336–37 (Reyna, J., dissenting); see also Kisor, 995 F.3d at 1372 (O'Malley, J., dissenting from denial of rehearing en banc) (advocating that the canon applies “[w]here differing plausible, reasonable interpretations of the terms of a regulation are possible” (emphasis added)). As I've said at length elsewhere, we shouldn't apply the pro-veteran canon “merely because a veteran-friendly construction is possible.” Prost Kisor Concurrence, 995 F.3d at 1355.
With those important preliminaries out of the way, I conclude for the reasons below that the pro-veteran canon does not preclude Auer deference here because (A) it is not one of the “traditional tools” we must apply before deferring under Auer and (B) resolution of the ambiguity here is best understood as delegated to the agency, not the courts.
In view of the Supreme Court's instruction that “a court must exhaust all the ‘traditional tools’ of construction” before finding ambiguity and deferring under Auer, Kisor, 139 S. Ct. at 2415 (cleaned up), I must first consider the suggestion that the pro-veteran canon is “one such traditional tool,” Pet'rs' Br. 35, 44 n.10 (quoting Procopio v. Wilkie, 913 F.3d 1371, 1383 (Fed. Cir. 2019) (en banc) (O'Malley, J., concurring)); see also Kisor v. Shulkin, 880 F.3d 1378, 1381 (Fed. Cir. 2018) (O'Malley, J., dissenting from denial of rehearing en banc) (“The [pro-veteran canon] is one of those rules of statutory construction.”). I conclude it is not.
Although the Court has not expressly disabused us of the notion that the pro-veteran canon is one of the “traditional tools,” it has made clear that not all canons fall within this category.4 One example is a similar tiebreaking rule, the rule of lenity. In Yates v. United States, the Court stated that “[f]inally, if our recourse to traditional tools of statutory construction leaves any doubt ․, we would invoke the rule that ambiguity concerning the ambit of criminal statutes should be resolved in favor of lenity.” 574 U.S. 528, 547–48, 135 S.Ct. 1074, 191 L.Ed.2d 64 (2015) (cleaned up). And more recently in Shular v. United States, the Court explained that lenity “applies only when, after consulting traditional canons of statutory construction, we are left with an ambiguous statute.” ––– U.S. ––––, 140 S. Ct. 779, 787, 206 L.Ed.2d 81 (2020). As I've observed, the Supreme Court's invocation of the pro-veteran canon across its “relatively short” history—typically just to “further confirm an interpretation that [the Court] reached by analyzing text and context”—resembles the rule of lenity and reflects this same order of operations. See Prost Kisor Concurrence, 995 F.3d at 1350–55.5
The pro-veteran canon's normative character also supports my conclusion that it is not one of the contemplated “traditional tools.” Descriptive canons, like the traditional tools mentioned in Kisor, “are just specific applications of the basic goal of interpretation: finding the ordinary meaning of statutory text.” Arangure v. Whitaker, 911 F.3d 333, 340 (6th Cir. 2018). But the pro-veteran canon is a substantive canon that “enter[s] the calculus when judges ‘need some way to finish the job and to pick from among the possible meanings that their primary interpretive tools have identified.’ ” Prost Kisor Concurrence, 995 F.3d at 1350 (quoting Caleb Nelson, What Is Textualism?, 91 Va. L. Rev. 347, 394 (2005)); see also Vitol, Inc. v. United States, 30 F.4th 248, 253 (5th Cir. 2022) (“[O]nly ․ after plain meaning and application of the interpretive canons are found lacking ․ do the so-called substantive canons ․ come into play.” (cleaned up)). They “express the law's supposed preferences when certain close interpretive calls arise.” Calderon v. Sixt Rent a Car, LLC, 5 F.4th 1204, 1219 (11th Cir. 2021) (Newsom, J., concurring). Accordingly, “[a]mbiguity canons merely instruct courts on how to ‘choos[e] between equally plausible interpretations of ambiguous text.’ ” W. Virginia v. EPA, ––– U.S. ––––, 142 S. Ct. 2587, 2620 n.3, 213 L.Ed.2d 896 (2022) (Gorsuch, J., concurring) (second alteration in original) (quoting Amy Coney Barrett, Substantive Canons and Faithful Agency, 90 B.U. L. Rev. 109, 109 (2010)). Any other role makes no sense “as a logical matter.” Prost Kisor Concurrence, 995 F.3d at 1354. “[I]f ‘interpretive doubt’ is a precondition for applying the canon,” then “the existence of interpretive doubt must be determined without employing the canon.” Id. “Otherwise, circularity results.” Id.
For these reasons, the pro-veteran canon is not one of the “traditional tools” we must apply before concluding that a regulation is ambiguous. Rather, it kicks in after a conclusion of ambiguity—similar to the way lenity, Chevron, Auer, and other ambiguity-triggered rules do.6
Having concluded that the pro-veteran canon need not be applied before finding ambiguity and applying Auer, I also conclude that Auer provides the proper resolution of the ambiguity in this case. In my view, deciding which rule applies here requires recognizing that “ambiguity is essentially a delegation of policymaking authority to the governmental actor charged with interpreting” the relevant provision, and that the pro-veteran canon and Auer deference are each a type of delegation—just to different actors. See Barrett, supra, at 123. Like Chevron deference, which “treat[s] statutory ambiguity as a delegation of gap-filling authority to an administrative agency,” id. at 182 n.66, Auer deference is similarly “rooted in a presumption about congressional intent,” i.e., “that Congress would generally want the agency to play the primary role in resolving regulatory ambiguities,” Kisor, 139 S. Ct. at 2412, as well as an “awareness that resolving genuine regulatory ambiguities often entails the exercise of judgment grounded in policy concerns,” id. at 2413 (cleaned up). “[T]he core theory of Auer deference is that sometimes the law runs out, and policy-laden choice is what is left over.” Id. at 2415.
In a similar vein, “a judge applying a canon like lenity”—or the pro-veteran canon—“to implement unclear text is not deviating from her best understanding of Congress's instructions”—that “Congress left the problem to her.” See Barrett, supra, at 123. Accordingly, “we presume[ ] Congress intended to invest interpretive power in whichever actor was best positioned to develop expertise about the given problem,” Kisor, 139 S. Ct. at 2417 (cleaned up), including “as between agencies and courts,” id.
Here, the ambiguity is in a gap-filling regulation on a matter well within the agency's expertise: the coverage of a medical diagnostic code. See id. at 2413 (“Agencies (unlike courts) have unique expertise, often of a scientific or technical nature, relevant to applying a regulation to complex or changing circumstances.” (cleaned up)). Specifically, DC 5055 sets forth ratings for knee replacements as a subsection of 38 C.F.R. § 4.71a. That regulation, in turn, is authorized by 38 U.S.C. § 1155, which broadly instructs that “[t]he Secretary shall adopt and apply a schedule of ratings of reductions in earning capacity from specific injuries or combination of injuries.” And the Secretary's explanation for his interpretation is “[t]he progression of this area of medical science.” J.A. 1.7
Conceivably, some regulatory ambiguities might be better resolved by courts applying the pro-veteran canon. Perhaps, for example, “[w]hen the agency has no comparative expertise in resolving a regulatory ambiguity,” since in such a case “Congress presumably would not grant it that authority.” Kisor, 139 S. Ct. at 2417 (indicating that “[s]ome interpretive issues,” like “the elucidation of a simple common-law property term” or “the award of an attorney's fee,” “may fall more naturally into a judge's bailiwick”). But that's not what we have here. Rather, the most logical conclusion is that Congress delegated resolution of the ambiguity here to the agency, which must “fill out the regulatory scheme Congress has placed under its supervision.” Id. at 2418. Indeed, we've deferred for similar reasons in the closely related Chevron context. In Buffington v. McDonough, for example, we did so where the Secretary “filled a statutory gap.” 7 F.4th 1361, 1367 (Fed. Cir. 2021). There, we also declined to apply the pro-veteran canon. Compare id. at 1366 n.5, with id. at 1374–75 (O'Malley, J., dissenting). Likewise in Terry v. Principi we deferred under Chevron to fill a gap instead of applying the pro-veteran canon. 340 F.3d 1378, 1383 (Fed. Cir. 2003). Accordingly, the ambiguity in this case implicates the rationales underlying Auer, not the pro-veteran canon.
The Guidance we review here is a forward-looking prescription, not a post-hoc rationalization or a reliance-upending departure from a prior agency position. With Auer still on the books after Kisor, we must defer to it absent any other applicable exceptions. Nothing in Hudgens prevents us from doing so in this case. I respectfully dissent.
1. The VA also included an exception to this interpretation for DC 5054, which relates to hip replacements. Knee Replacement Guidance, 80 Fed. Reg. at 42,041–42. That exception is not relevant here.
2. Specifically, we have jurisdiction over the amended petition because either the Knee Replacement Guidance or the Knee Replacement Manual Provision is reviewable under § 502, regardless of which is the reviewable interpretive rule under § 552(a)(1). NOVA 2020, 981 F.3d at 1382.
3. The dissent concludes that the Guidance does not upend reliance because it applies prospectively and not retrospectively. Diss. at 3. But the dissent does not identify cases where courts have relied on a distinction between prospective and retrospective conflicting agency interpretations. Instead, the Supreme Court has held that the “disruption of expectations may occur when an agency substitutes one view of a rule for another.” Kisor, 139 S Ct. at 2418. An agency interpretation is not entitled to Auer deference when the interpretation “does not reflect the agency's fair and considered judgment,” which may occur “when the agency's interpretation conflicts with a prior interpretation” or is a “post hoc rationalization.” Christopher, 567 U.S. at 155, 132 S.Ct. 2156. Thus, courts have not given Auer deference to conflicting agency interpretations that are applied in subsequent proceedings. See, e.g., Romero, 937 F.3d at 296–97 (declining to give Auer deference to precedential decision of Attorney General that Board of Immigration Appeals applied to end administrative closure of case); Rafferty v. Denny's, Inc., 13 F.4th 1166, 1187 (11th Cir. 2021) (declining to give Auer deference to agency Opinion Letter issued prior to filing of complaint at issue). The Guidance, like the Board decision in Hudgens, conflicts with prior interpretations, which forecloses Auer deference. See Hudgens, 823 F.3d at 637–38 & n.3.
4. “This court has not definitively resolved at what stage the pro-veteran canon applies and whether it precedes any claims of deference to an agency interpretation.” Roby v. McDonough, 2021 WL 3378834, at *8 (Fed. Cir. Aug. 4, 2021). Because we conclude that the Secretary's interpretation is not entitled to Auer deference, we decline to opine on whether the pro-veteran canon precedes or follows Auer deference.
1. Notably, in view of Kisor v. Wilkie, “Board decisions ․ appear not to be entitled to Auer deference.” Nat'l Org. of Veterans' Advocs., Inc. v. Sec'y of Veterans Affs., 981 F.3d 1360, 1382 n.14 (Fed. Cir. 2020) (en banc) (citing ––– U.S. ––––, 139 S. Ct. 2400, 2424, 204 L.Ed.2d 841 (2019) (stating that the Solicitor General suggested Auer deference may be inappropriate for Board decisions)). True, that's not the issue here, Maj. ––––, but Kisor at minimum provides reason to question the “weight” Hudgens assigned those decisions. Hudgens, 823 F.3d at 638.
3. The Supreme Court subsequently declined the invitation to address “the appropriate use of the pro-veteran canon.” See Petition for Writ of Certiorari at 11, Kisor, 142 S. Ct. 756.
4. Kisor repeatedly describes the “traditional tools” as at least including “text, structure, history, and purpose.” E.g., 139 S. Ct. at 2415. The Court's omission of the pro-veteran canon from this list might be unremarkable except that one of the two questions presented in Kisor at the certiorari stage was whether Auer should yield to the pro-veteran canon. The Court did not grant certiorari on that question.
5. A different (and more dubious) conception of the pro-veteran canon is not as a tiebreaking rule, but as a broad liberal-construction principle. Prost Kisor Concurrence, 995 F.3d at 1351–54. This view ignores that “[l]egislation is ․ the art of compromise, the limitations expressed in statutory terms [are] often the price of passage, and no statute yet known pursues its stated purpose at all costs.” Henson v. Santander Consumer USA Inc., ––– U.S. ––––, 137 S. Ct. 1718, 1725, 198 L.Ed.2d 177 (2017) (cleaned up).
6. Concluding otherwise would mean “the VA, alone among the executive agencies, is not entitled to deference in interpreting its regulations,” which “would be anomalous to say the least.” Kisor, 995 F.3d at 1361 (Hughes, J., concurring in denial of rehearing en banc).
7. “The field of orthopedic medicine has progressed to such a degree that total prosthetic replacement of a joint is not always necessary․ Partial replacement has the benefit of not requiring the same length of time for convalescence. The progression of this area of medical science has raised an issue as to whether a veteran who undergoes a partial replacement of a joint is entitled to the 100-percent rating evaluation during the convalescent period under DCs 5051 through 5056.” J.A. 1 (footnote omitted).
Cunningham, Circuit Judge.
Dissenting opinion filed by Circuit Judge Prost.