Philbrick v. Azar II

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

SAMUEL PHILBRICK, et al.,

Plaintiffs, v. Civil Action No. 19-773 (JEB)

ALEX M. AZAR II, et al.,

Defendants.

MEMORANDUM OPINION

In November 2018, the Secretary of Health and Human Services approved the State of New Hampshire’s proposal to impose work requirements on a significant share of its Medicaid recipients. Under the proposal, most non-disabled Medicaid beneficiaries ages 19 to 64 would be required to demonstrate that they have completed 100 hours of qualifying employment or other “community-engagement” activities each month (or show that they satisfy an exemption) or risk losing their health-care coverage. Four New Hampshire residents have challenged the Secretary’s approval in this Court, arguing that it violates the Administrative Procedure Act and the Constitution.

The issues presented in this case are all too familiar. In the past year or so, this Court has resolved challenges to similar programs in Kentucky and Arkansas, each time finding the Secretary’s approval deficient. See Stewart v. Azar, 366 F. Supp. 3d 125, 131 (D.D.C. 2019) (Stewart II); Gresham v. Azar, 363 F. Supp. 3d 165, 169 (D.D.C. 2019); Stewart v. Azar, 313 F. Supp. 3d 237, 243 (D.D.C. 2018) (Stewart I). The overriding shortcoming in the agency’s decisions in those cases was its failure to adequately consider the requirements’ effects on Medicaid coverage. Despite conceding that providing medical care to the needy is “Medicaid’s

2

core objective,” Gresham, 363 F. Supp. 3d at 176 (citation omitted), HHS did not “offer its own estimates of coverage loss or grapple with comments in the administrative record projecting that the proposal would lead a substantial number of residents to be disenrolled from Medicaid.” Id. at 175 (cleaned up).

Plaintiffs argue that the Secretary’s approval of New Hampshire’s plan suffers from the same deficiency and thus must meet the same fate. The Court concurs. On their face, these work requirements are more exacting than Kentucky’s and Arkansas’s, mandating 100 monthly hours — as opposed to 80 — of employment or other qualifying activities. They also encompass a larger age range than in Arkansas, which applied the requirements only to persons 19 to 49. Yet the agency has still not contended with the possibility that the project would cause a substantial number of persons to lose their health-care coverage. That omission is particularly startling in light of information before the Secretary about the initial effects of Arkansas’s markedly similar project — namely, that more than 80% of persons subject to the requirements had reported no compliance information for the initial months, and nearly 16,900 people had lost coverage. The agency’s rejoinders — that the requirements advance other asserted purposes of Medicaid, such as the health and financial independence of beneficiaries and the fiscal sustainability of the safety net — are identical to those this Court rejected with respect to HHS’s 2018 approval of Kentucky’s program. Perhaps seeing the writing on the wall, the Government conceded at oral argument that its reasoning was deficient in these respects under the analysis in the Court’s prior Opinions.

In short, we have all seen this movie before. The Secretary has significant discretion to approve demonstration projects that promote the objectives of the Medicaid Act, and it is not for the Court to second guess his policy decisions or substitute its judgment for his. “But courts

3

retain a role, and an important one, in ensuring that agencies have engaged in reasoned decisionmaking.” Judulang v. Holder, 565 U.S. 42, 53 (2011). At the heart of this review is an assessment of “whether the decision was based on a consideration of the relevant factors.” Id. (citation omitted). For the fourth time, HHS has fallen short of this fundamental administrative- law requirement. The Court will, accordingly, grant summary judgment to Plaintiffs and vacate the Secretary’s approval of New Hampshire’s community-engagement requirements. I. BACKGROUND The Court begins with a now-familiar overview of the relevant history and provisions of the Medicaid Act. It then turns to New Hampshire’s challenged plan before concluding with the procedural history of this case.

The Medicaid Act

Since 1965, the federal government and the states have worked together to provide medical assistance to certain vulnerable populations under Title XIX of the Social Security Act, commonly known as Medicaid. See 42 U.S.C. § 1396-1. The Centers for Medicare and Medicaid Services (CMS), a federal agency within the Department of Health and Human Services, has primary responsibility for overseeing Medicaid programs. Under the cooperative federal-state arrangement, participating states submit their “plans for medical assistance” to the Secretary of HHS. Id. To receive federal funding, those plans — along with any material changes to them — must be “approved by the Secretary.” Id.; see also 42 C.F.R. § 430.12(c). Currently, all states have chosen to participate in the program.

To be approved, state plans must comply with certain minimum parameters set out in the Medicaid Act. See 42 U.S.C. § 1396a (listing 86 separate requirements). One such provision requires state plans to “mak[e] medical assistance available” to certain low-income individuals.

4

Id. § 1396a(a)(10)(A). Until recently, that group included pregnant women, children, and their families; some foster children; the elderly; and people with certain disabilities. Id. In 2010, however, Congress enacted the Patient Protection and Affordable Care Act (ACA), colloquially known as Obamacare, “to increase the number of Americans covered by health insurance.” Nat’l Fed’n of Indep. Business v. Sebelius, 567 U.S. 519, 538 (2012). Of relevance here, that statute required participating states to expand Medicaid coverage to additional low-income adults under 65 who did not previously qualify. See 42 U.S.C. § 1396a(a)(10)(A)(i)(VIII).

Generally, a state must cover all qualified individuals or forfeit its federal Medicaid funding. Id. § 1396a(a)(10)(B); id. § 1396c. That was originally so for the ACA expansion population as well. Id. § 1396c. In NFIB, however, the Supreme Court held that Congress could not, consistent with the Spending Clause of the Constitution, condition a state’s entire Medicaid funds on its agreeing to the expansion. See 567 U.S. at 584–85. As a result, states could choose not to cover the new population and lose no more than the funds that would have been appropriated for that group. Id. at 587. If the state, conversely, does decide to provide coverage, those individuals would become part of its mandatory population. Id. at 585–87 (explaining that Congress may “offer[] funds under the Affordable Care Act to expand the availability of health care, and require[] that States accepting such funds comply with the conditions on their use”). In that instance, the state must afford the expansion group “full benefits” — i.e., it must provide “medical assistance for all services covered under the State plan” that are substantially equivalent “in amount, duration, or scope . . . to the medical assistance available for [other] individual[s]” covered under the Act. See 42 U.S.C. § 1396d(y)(2)(B); 42 C.F.R. § 433.204(a)(2).

5

The Medicaid Act, in addition to defining who is entitled to coverage, also ensures what coverage those enrolled individuals receive. Under § 1396a, states must cover certain basic medical services, see 42 U.S.C. §§ 1396a(a)(10)(A), 1396d(a), and the statute limits the amount and type of premiums, deductions, or other cost-sharing charges that a state can impose on such care. Id. § 1396a(a)(14); see also id. § 1396o. Other provisions require states to provide up to three months of retroactive coverage once a beneficiary enrolls, id. § 1396a(a)(34), and to ensure that recipients receive all “necessary transportation . . . to and from providers.” 42 C.F.R. § 431.53. Finally, states must “provide such safeguards as may be necessary to assure that eligibility” and services “will be provided, in a manner consistent with simplicity of administration and the best interests of the recipients.” 42 U.S.C. § 1396a(a)(19).

Both before and after the passage of the ACA, a state accepting federal Medicaid funds is not entirely locked in; instead, if it wishes to deviate from certain of the Act’s requirements, it can seek a waiver from the Secretary of HHS. See 42 U.S.C. § 1315. In particular, Section 1115 of the Social Security Act allows the Secretary to approve “experimental, pilot, or demonstration project[s] which, in [his] judgment . . . , [are] likely to assist in promoting the [Act’s] objectives.” 42 U.S.C. § 1315(a). As conceived, experimental projects were “expected to be selectively approved by the Department and to be those which are designed to improve the techniques of administering assistance.” S. Rep. No. 1589, 87th Cong., 2d Sess. 19, reprinted in 1962 U.S.C.C.A.N. 1943, 1962. Once the Secretary has greenlighted such a project, he can then waive compliance with the requirements of § 1396a “to the extent and for the period . . . necessary to enable [the] State . . . to carry out such project.” Id. § 1315(a)(1).

While the ultimate decision whether to grant § 1115 approval rests with the Secretary, his discretion is not boundless. Before HHS can act on a waiver application, the state “must provide

6

at least a 30-day public notice[-]and[-]comment period” regarding the proposed program and hold at least two hearings at least 20 days before submitting the application. See 42 C.F.R. §§ 431.408(a)(1), (3). Once a state completes those prerequisites, it then sends an application to CMS. Id. § 431.412 (listing application requirements). After the agency notifies the state that it has received the waiver application, a federal 30-day public-notice period commences, and the agency must wait at least 45 days before rendering a final decision. Id. §§ 431.416(b), (e)(1).

Factual Background

1. New Hampshire Granite Advantage In 2014, New Hampshire, like many states, expanded Medicaid under the ACA to previously uninsured adults whose income is 133 percent of the federal poverty line or less. See AR 17; AR 1949. More than 53,000 individuals have received coverage as a result, helping to reduce the State’s uninsured rate by 45 percent. Id. at 4384. Since 2015, the State has covered this population through Section 1115 demonstration projects that deviate from traditional Medicaid delivery mechanisms — first adopting a premium-assistance model and later shifting to a managed-care system. Id. at 4379. While New Hampshire has had an interest in work requirements dating back to 2016, id. at 99, it proposed to amend its demonstration to add the work and community-engagement requirements under consideration in this suit in 2018. Id. at 4377.

As proposed, the project — now called Granite Advantage — requires most non-disabled adults aged 19 to 64 to complete 100 hours per month of employment or other community activities. Id. at 4. Certain categories of beneficiaries are exempt, including caregivers for a dependent child, pregnant women, and the medically frail. Id. at 5. If a beneficiary does not demonstrate compliance with the work requirements in a particular month, she will be sent a

7

notice stating that her Medicaid will be terminated the following month if she does not make up the hours or show that she qualifies for an exemption. Id. Once a beneficiary’s coverage is suspended, it can be reactivated by completing 100 hours of qualifying activities or obtaining an exemption. Id. at 5, 7. Separately, New Hampshire requested as part of these amendments that HHS allow the State to eliminate all retroactive coverage. Id. at 4377.

The Secretary approved the amendments on November 30, 2018, explaining that they promoted the purposes of the Medicaid Act because they would improve the “health and wellness” of beneficiaries and enhance the “fiscal sustainability of the Medicaid program.” Id. at 1–2. With respect to commenters’ concerns that some beneficiaries would lose coverage, the agency responded that “the demonstration will provide coverage to individuals that the state is not required to cover” — namely, the ACA expansion population. Id. at 10. Indeed, because “the state plans to end its current coverage of the new adult group” in the event the project were not approved, HHS says, Granite Advantage necessarily increases coverage. Id. at 6, 10. The agency further explained that the requirements were “not designed to encourage” coverage loss and are “intended to [be] achievable,” citing certain exemptions and safeguards that are meant to reduce the likelihood of persons improperly losing their Medicaid. Id. at 10–11.

While the new requirements could have been implemented under this approval beginning January 1, 2019, id. at 1, they have still not been put into full effect. New Hampshire, after several initial delays, required beneficiaries to submit qualifying hours or proof of an exemption this past June. See ECF No. 1 (Complaint), ¶ 10. Under that timeframe, persons who did not satisfy the reporting obligations would lose their coverage on August 1. Id. As of July 8, 2019, however, approximately 17,000 non-exempt beneficiaries (out of about 25,000 total) had not reported any compliance information to the New Hampshire Department of Health and Human

8

Services. See ECF No. 44-2 (Jeffrey A. Meyers Letter, July 8, 2019) at 3. Citing this consideration and emphasizing the difficulty the State has had in communicating with persons subject to the community-engagement requirements, the Department announced that it was further delaying implementation until September 30, 2019. See ECF No. 44 (Notice) at 2. Under the new implementation plan, Medicaid beneficiaries who do not report compliance with the requirements would lose coverage beginning December 1. Around the same time, the New Hampshire Legislature amended the program in several respects, including by expanding the scope of the exemptions. The State explained that it plans to seek reapproval of such amendments from CMS over the next several months. See Oral Argument Transcript (Provisional) at 3, 17.

2. Other CMS Approvals

New Hampshire is not the only state that has been interested in work requirements. As noted at the start, CMS has approved similar proposals submitted by Kentucky and Arkansas, each of which has been challenged and struck down in this Court. Kentucky’s program — called Kentucky HEALTH — mirrors New Hampshire’s in many respects. As relevant here, it requires non-exempt adults aged 19 to 64 who receive coverage through the expansion to complete and report 80 hours per month of qualifying activities, such as employment, education, or job training. See Stewart I, 313 F. Supp. 3d at 246. The failure to do so or to report an exemption results in the termination of Medicaid coverage. Id. at 246–47.

Before the requirements took effect in the Commonwealth, several Medicaid recipients sought judicial review of HHS’s approval. Id. at 248. They argued, among other things, that the agency had failed to adequately explain why Kentucky HEALTH promoted the objectives of Medicaid and that the approval of the project exceeded the Secretary’s statutory authority. The

9

Court agreed with Plaintiffs in one central and dispositive respect: “[T]he Secretary never adequately considered whether Kentucky HEALTH would in fact help the state furnish medical assistance to its citizens, a central objective of Medicaid.” Id. at 243. It therefore vacated the Secretary’s approval and remanded the matter to the agency for further consideration. Id. at 273–74. HHS subsequently reopened the comment period and reapproved Kentucky’s project on November 20, 2018. The agency reasoned, along substantially similar lines as it did ten days later when it approved New Hampshire’s project, that Kentucky HEALTH advanced the Medicaid Act’s objectives because it would 1) promote the health and financial independence of beneficiaries, a justification the Court had found wanting in the first round, 2) increase coverage because it allows Kentucky to cover the expansion population when it would not do so otherwise, and 3) advance the fiscal sustainability of the state’s Medicaid program. See Stewart II, 366 F. Supp. 3d at 138. Believing these justifications still unsatisfactory, the Bluegrass State plaintiffs returned to this Court, which again concurred. Concluding that the agency’s previous rationales fared no better and that its new explanation still failed to grapple with the possibility of coverage loss, the Court vacated the approval. Id. at 138–39.

Arkansas’s project, named the Arkansas Works Amendments, followed a similar, although abbreviated, path. The State proposed to require most able-bodied beneficiaries in the expansion population aged 19 to 49 to complete 80 hours of qualifying employment or other activities. See Gresham, 363 F. Supp. 3d at 172. Non-exempt individuals who did not report sufficient qualifying hours for three consecutive months in a calendar year would be disenrolled from Medicaid for the remainder of that year. Id. The Secretary approved the requirements on March 5, 2018, and their roll-out was staged through 2018 and early 2019. Id. During the first six months after implementation, however, “only a small percentage of the persons required to

10

report compliance . . . actually did so” — in October 2018, only 12.3% reported any kind of qualifying activities — and more than 16,900 persons lost Medicaid coverage for some period of time as a result. Id. Several beneficiaries challenged the program under the APA, and, finding the Secretary’s explanation deficient for the same reasons as in its first Kentucky decision, the Court vacated his approval. Id. at 175. The Court’s decisions as to both of those cases are now on appeal before the D.C. Circuit. See Case Nos. 19-5094, 19-5095, 19-5096, 19-5097. No oral argument date has yet been set.

Procedural History

Believing with Shakespeare that what’s past is prologue, four New Hampshire residents filed this lawsuit on March 20, 2019. Like the plaintiffs in Arkansas and Kentucky, they assert that the Secretary’s approval of the proposed community-engagement requirements violates the APA and the Constitution. Because it was designated as related to Stewart and Gresham, the case was directed to this Court. See ECF Nos. 2–3. The State of New Hampshire has since intervened as a Defendant, and numerous amici have also weighed in. Dueling Cross-Motions for Summary Judgment are now ripe, and the Court held a hearing on July 23, 2019. II. LEGAL STANDARD The parties have cross-moved for summary judgment on the administrative record. The summary-judgment standard set forth in Federal Rule of Civil Procedure 56(c), therefore, “does not apply because of the limited role of a court in reviewing the administrative record.” Sierra Club v. Mainella, 459 F. Supp. 2d 76, 89 (D.D.C. 2006); see also Bloch v. Powell, 227 F. Supp. 2d 25, 30 (D.D.C. 2002), aff’d, 348 F.3d 1060 (D.C. Cir. 2003). “[T]he function of the district court is to determine whether or not as a matter of law the evidence in the administrative record permitted the agency to make the decision it did.” Sierra Club, 459 F. Supp. 2d. at 90 (quotation

11

marks and citation omitted). “Summary judgment is the proper mechanism for deciding, as a matter of law, whether an agency action is supported by the administrative record and consistent with the [Administrative Procedure Act] standard of review.” Loma Linda Univ. Med. Ctr. v. Sebelius, 684 F. Supp. 2d 42, 52 (D.D.C. 2010) (citation omitted).

The Administrative Procedure Act “sets forth the full extent of judicial authority to review executive agency action for procedural correctness.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 513 (2009). It requires courts to “hold unlawful and set aside agency action, findings, and conclusions” that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2). Agency action is arbitrary and capricious if, for example, the agency “entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983).

In other words, an agency is required to “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” Id. at 43 (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168 (1962)) (internal quotation marks omitted). Courts, accordingly, “do not defer to the agency’s conclusory or unsupported suppositions,” United Techs. Corp. v. Dep’t of Def., 601 F.3d 557, 562 (D.C. Cir. 2010) (quoting McDonnell Douglas Corp. v. Dep’t of the Air Force, 375 F.3d 1182, 1187 (D.C. Cir. 2004)), and “agency ‘litigating positions’ are not entitled to deference when they are merely [agency] counsel’s ‘post hoc rationalizations’ for agency action, advanced for the first time in the reviewing court.” Martin v. Occupational Safety & Health

12

Review Comm’n, 499 U.S. 144, 156 (1991) (citation omitted). Although a reviewing court “may not supply a reasoned basis for the agency’s action that the agency itself has not given,” a decision that is not fully explained may nevertheless be upheld “if the agency’s path may reasonably be discerned.” Bowman Transp., Inc. v. Arkansas-Best Freight System, Inc., 419 U.S. 281, 285–86 (1974) (citation omitted). III. ANALYSIS Just as their predecessors did, Plaintiffs here challenge the Secretary’s approval of New Hampshire’s demonstration project on a number of different grounds, including that it was arbitrary and capricious, in excess of statutory authority, and in violation of the Take Care Clause of the U.S. Constitution. As in its three previous Opinions, the Court need only address the first to resolve this case. Before turning to that issue, however, it begins with jurisdiction.

Jurisdiction

While Defendants largely do not contest whether subject-matter jurisdiction exists in this case, the Court has an independent duty to assure as much. See Steel Co. v. Citizens for a Better Environment, 523 U.S. 83, 94–95 (1998). Two potential hurdles stand across Plaintiffs’ path. The first is whether the Secretary’s decision is “committed to agency discretion by law” and is therefore unreviewable under the APA. See 5 U.S.C. § 701(a)(2). HHS insisted in Stewart I that its determination to approve a Section 1115 demonstration project fell within this exception to the general presumption that administrative action is judicially reviewable. After a lengthy discussion, the Court agreed with “every court which has considered the issue” and found that the Secretary’s approval was “subject to APA review.” 313 F. Supp. 3d at 256 (quoting Beno v. Shalala, 30 F.3d 1057, 1067 & n.24 (9th Cir. 1994)). Apart from a couple of conclusory

13

sentences in their Opposition and Reply, Defendants no longer press this objection, and the Court sees no reason to depart from its prior decision finding HHS’s decision reviewable.

The second jurisdictional question is whether Plaintiffs have standing to bring this suit.

Article III of the Constitution limits the Court’s jurisdiction to “cases” or “controversies.” The standing doctrine enforces this requirement, assuring that courts only decide actual disputes between parties with personal stakes in the outcome. See Clinton v. City of New York, 524 U.S. 417, 429 (1998). To establish standing, Plaintiffs must show that they have suffered (or will suffer in the future) a concrete injury that is both fairly traceable to the challenged conduct and likely to be redressed by a favorable judicial decision. See Lujan v. Defenders of Wildlife, 504 U.S. 555, 590 (1992). While Defendants disputed the basis of the plaintiffs’ standing in Stewart I, in their briefing this time around, they generally leave their powder dry. Upon review, the Court can see why: at least one Plaintiff clearly has standing.

Only a few sentences are needed to show as much. Plaintiff Ian Ludders is a forty-year-

old Medicaid recipient who will be subject to New Hampshire’s community-engagement requirements. See ECF No. 1 (Compl.), ¶¶ 146–56; ECF No. 19-3 (Declaration of Ian Ludders); see also ECF 37-2 (Declaration of Henry Lipman) at 2. Because he maintains a “subsistence lifestyle that prioritizes living off the land” and his work is agricultural and thus largely seasonal, he does not expect that he will be able to comply with the new requirements for multiple months out of the year. See Compl., ¶¶ 147–56. Taken together, these facts satisfy each of the three standing requirements: 1) there is a substantial risk that Ludders will lose his Medicaid coverage, thereby injuring him; 2) this risk is traceable to the Secretary’s approval of the requirements; and 3) a judicial decision vacating them is likely to prevent the future injury. It may well be that Plaintiffs apart from Ludders also have standing, but there is no need to delve into that issue

14

here. See Comcast Corp. v. FCC, 579 F.3d 1, 6 (D.C. Cir. 2009) (“[I]f one party has standing in an action, a court need not reach the issue of the standing of other parties when it makes no difference to the merits of the case.”).

Defendants separately suggest that none of the Plaintiffs has standing to challenge one aspect of the Secretary’s approval: his decision to allow New Hampshire to eliminate retroactive Medicaid coverage. See ECF No. 30 (HHS Cross-Mot.) at 24. As the Court explained in its prior Opinions, however, it is appropriate to “examine[] the approval of the project as a whole,” given the nature of Plaintiffs’ claim and the administrative action under review. Stewart I, 313 F. Supp. 3d at 253. That makes it unnecessary to decide whether any Plaintiff has standing with respect to this particular component of the Secretary’s approval.

Merits

Turning to the merits, Plaintiffs principally submit that the Secretary’s approval of the Granite Advantage project is arbitrary and capricious because it did not adequately consider the effects of the demonstration project on Medicaid coverage. The Court, as discussed above, found this argument persuasive in each of its prior three decisions, and it continues to do so. The critical issues are thus whether the reasoning in the Secretary’s approval letter meaningfully differs from the previous three, or whether attributes particular to New Hampshire’s community- engagement program, as discussed in the approval, suggest coverage-loss concerns will be less significant.

Defendants’ briefing does not attempt to distinguish the approval letter or the program from those in Stewart I, Stewart II, and Gresham; indeed, it marches through its arguments barely acknowledging that the Court has decided these precise issues before and adversely to HHS. At oral argument, however, the Government effectively conceded that the Secretary’s

15

reasoning in this case cannot be distinguished from his explanations in the prior ones. See Tr. at 2, 10. The Court likewise finds the records to be indistinguishable. As discussed in more detail below, CMS’s approval letter mirrors the one in Stewart II, with numerous key paragraphs matching it word for word. And New Hampshire’s proposed project presents, if anything, greater coverage-loss concerns than Kentucky’s and Arkansas’s, given the hours requirement and the age range to whom it applies.

The Court’s analysis unfolds in two parts. First, it summarizes the now-familiar view that the core objective of the Medicaid Act is to furnish health-care coverage to the needy and explains why the Secretary failed to adequately consider that objective here. Turning to Defendants’ counterarguments — nearly all of which were addressed at length in Stewart II — the Court will offer an abbreviated restatement of why the agency’s consideration of other Medicaid Act objectives does not remedy this deficiency.

1. Coverage as Objective of the Medicaid Act The Secretary, as outlined above, can only approve demonstration projects that are “likely to assist in promoting the objectives” of the Medicaid Act. See 42 U.S.C. § 1315(a). Before greenlighting a project, he must therefore identify the objectives of the Act and explain why the demonstration is likely to promote them. The Court has assumed that the Secretary’s interpretation of those objectives is entitled to Chevron deference. See Gresham, 363 F. Supp. 3d at 176. That is, in reviewing his understanding, the Court must first ask whether “Congress has directly spoken to the precise question at issue” and, if not, whether “the agency’s answer is based on a permissible construction of the statute.” Chevron U.S.A., Inc. v. Nat’l Res. Def. Council, Inc., 467 U.S. 837, 842–43 (1984). Because all parties agree that “Medicaid’s core objective” is to “furnish[] medical assistance” to persons who cannot afford it, according such

16

deference would have no practical significance with respect to this objective. See ECF No. 38 (HHS Reply) at 5; ECF No. 35 (Plaintiffs’ Reply) at 6; see also Stewart I, 313 F. Supp. 3d at 260–62. The Court will not repeat its discussion of why the provision of medical assistance to beneficiaries — both recipients of traditional Medicaid and members of the expansion population — is the central purpose of the Act, but instead directs interested readers to its prior Opinions. See Gresham, 363 F. Supp. 3d at 176; Stewart II, 366 F. Supp. 3d at 138; Stewart I, 313 F. Supp. 3d at 260–62.

Having correctly identified the provision of Medicaid coverage as a core objective, the agency was required to reasonably explain whether New Hampshire’s proposed community- engagement requirements would advance or impede that goal. In other words, “the Secretary needed to consider whether the demonstration project would be likely to cause recipients to lose coverage and whether it could cause others to gain coverage.” Gresham, 363 F. Supp. 3d at 177. He once again neglected to do so on both counts.

a. Risk to Coverage

Before approving a proposed demonstration, the Secretary must address whether it creates a risk that beneficiaries will lose their Medicaid coverage. Unlike in Stewart I and Gresham, here the agency at least mentioned the possibility of coverage loss in its approval. That is step one. But “stating that a factor was considered . . . is not a substitute for considering it,” Getty v. Fed. Savs. & Loan Ins. Corp., 805 F.2d 1050, 1055 (D.C. Cir. 1986); he must “adequately analyze . . . the consequences” of his actions. Am. Wild Horse Pres. Campaign v. Perdue, 873 F.3d 914, 932 (D.C. Cir. 2017). Here the Secretary fell well short.

For starters, he “never provided a bottom-line estimate of how many people would lose Medicaid” with Granite Advantage in place. Stewart II, 366 F. Supp. 3d at 140 (quoting Stewart

17

I, 313 F. Supp. 3d at 262). In its proposal, New Hampshire estimated that the project would have no material effect on Medicaid enrollment, though it also hinted that coverage would otherwise have expanded given population growth. See AR 4386. The many commenters who addressed the issue unanimously agreed that coverage loss would be substantial. The Kaiser Family Foundation, for example, projected an enrollment loss between 6 and 17 percent, corresponding to between 2600 and 7500 people losing Medicaid. See AR 2532, 4384 (applying disenrollment rates to 44,000 non-frail adults subject to Granite Advantage). Appendix A documents numerous comments and studies projecting the same or greater coverage losses. See, e.g., AR 1949, 1953, 2132, 2208, 2210, 2238, 2241. When the Secretary approved the project, he also had the benefit of data from the first several months of Arkansas’s comparable project, which presented a stark picture. See AR 2731–47. In that time, only 12.3% of non-exempt persons “reported any kind of qualifying activity,” and “16,900 individuals ha[d] lost Medicaid coverage for some period of time” as a result. Gresham, 363 F. Supp. 3d at 172 (emphasis added). Commenters explained that results were likely to be similar in New Hampshire. See, e.g., AR 2563, 2963. Indeed, because the State’s requirements are more stringent than Arkansas’s in key respects — e.g., requiring 100 rather than 80 hours per month of activities and applying to adults aged 19 to 64 rather than 19 to 49 — many projected that Granite Advantage would lead to more coverage loss than Arkansas’s program. See, e.g., AR 2258, 2542, 2575, 2586.

What does the Secretary think about all this? Does he concur with New Hampshire’s apparent view that coverage loss is going to be minimal, or does he agree with the commenters that it is likely to be substantial? Are the coverage losses in Arkansas likely to be replicated in New Hampshire? We have no idea, since the approval letter offers no hints. While Defendants may well be correct that HHS does not need to provide a precise numeric estimate of coverage

18

loss, it can hardly be disputed that the agency needs to address the magnitude of that loss. That is particularly so where the comments uniformly assert — and the record evidence from similar programs strongly suggests — that the loss will be substantial. The Secretary’s “failure to address” this “salient factor” renders his decision arbitrary and capricious. See Humane Soc’y of United States v. Zinke, 865 F.3d 585, 606–07 (D.C. Cir. 2017).

HHS generally conceded at oral argument that the Secretary did not consider coverage to a greater degree in this case than in Stewart II. See Tr. at 10–11. In their briefing, however, Defendants offer several responses worthy of discussion. The agency first insists that it did consider the risk of coverage loss because it mentioned that possibility at several points in its approval letter. For example, the Secretary explained that “[t]he community engagement requirements may impact overall coverage levels if the individuals subject to the requirements choose not to comply with them.” AR 6. This and statements like it butter few parsnips, for acknowledging the possibility of coverage loss is not the same as analyzing that possibility. See Getty, 805 F.2d at 1055. The letter gives no indication, as mentioned above, about the seriousness of the loss; for all the reader gleans, the project could expel 75% of prior Medicaid beneficiaries. It also neglects the possibility, likewise discussed by numerous commenters, that many beneficiaries will lose coverage merely because they are not able to satisfy the administrative burden associated with reporting their compliance. See AR 1484, 1489, 2132, 2241, 3406, 4564.

The Secretary also offers a second, more substantive response to concerns about coverage loss. He says that Granite Health was designed “to minimize coverage loss due to noncompliance,” AR 11, citing the exemptions built into the project, the procedural safeguards intended to prevent improper disenrollment, and the otherwise achievable nature of the

19

requirements. Id. at 7–11. The State of New Hampshire likewise cites the broader intentions of the program and the corresponding protections put in place to prevent coverage loss. See ECF No. 37 (NH Reply) at 8. To state the obvious, however, that a project is intended to avoid coverage loss does not mean that it will do so. Similar intentions existed and corresponding protections were put in place in Kentucky, but the Commonwealth projected a coverage loss equivalent to 95,000 people losing Medicaid for one year. Same with Arkansas, yet it found that nearly 17,000 lost coverage at some point in the first six months alone. The commenters in this case explained as much, since most of the exemptions or safeguards CMS mentions were baked into their comments about the likelihood of significant coverage loss. See, e.g., AR 1484, 2132, 2241. Ultimately, the agency’s explanation here comes up short — just as it did in the previous three cases — because it does not address whether and how these design attributes bear on the actual magnitude of coverage loss. See Stewart II, 366 F. Supp. 3d at 142–43.

While Defendants offer two remaining arguments as to why they have adequately considered coverage, both are more properly addressed in subsequent sections. Briefly, they assert that the project will not actually decrease coverage because it allows the State to cover a population it would not otherwise cover — namely, the ACA expansion group. The Court will analyze this argument in the ensuing section on whether the Secretary adequately considered if the project would promote coverage. See infra Section III.B.1.b. Defendants also maintain that any coverage loss is outweighed by the project’s promotion of other purposes of Medicaid, including health, financial independence, and fiscal sustainability. This point — which does not directly bear on whether the Secretary has ever grappled with coverage — will be addressed below in Section III.B.2.

20

b. Promote Coverage

The Secretary offers two possible reasons why Granite Advantage might improve the state’s ability to furnish medical assistance to the needy. The first can be dispatched quickly. He says that the elimination of retroactive eligibility may “encourage beneficiaries to obtain and maintain health coverage, even when they are healthy.” AR 12. Setting aside the equivocal nature of this assertion and the numerous comments to the contrary in the record (AR 1479–80, 4565), this possibility has nothing to do with the coverage loss the community-engagement requirements might cause. To the extent the Secretary believes that the elimination of retroactive coverage might (counterintuitively) increase coverage, he needs to weigh the promotion side of the scale against the risk-of-loss side when approving the project. That did not happen. Indeed, such a calculus would be difficult to undertake given the agency’s failure, discussed at length above, to characterize the magnitude of coverage loss presented by the community-engagement requirements.

Defendants’ second argument about coverage promotion is one the Court addressed at length in Stewart II: because the State will “simply de-expand Medicaid” if Granite Advantage is not approved, any coverage provided to the expansion population through the demonstration is properly understood as increasing Medicaid coverage. See 366 F. Supp. 3d at 153; see also AR 10. In other words, “[a] demonstration that shrinks coverage may thus be coverage promoting for the purposes of § 1115 as long as the state threatens that if the demonstration is not approved, it will discontinue coverage entirely.” Stewart II, 366 F. Supp. 3d at 153; see HHS Cross-Mot. at 17–18. In Stewart II, the Court addressed this point as a variant of the agency’s fiscal- sustainability rationale because the ostensible reason for Kentucky to de-expand absent approval

21

was its fiscal situation. In doing so, however, the Court noted that the argument “does not depend on a state’s being in a fiscally precarious position because it does not take into account the reason the state wants to discontinue participating in the Medicaid program.” Stewart II, 366 F. Supp. 3d at 154.

This case proves the point. New Hampshire has not justified its desire to de-expand absent approval of the community-engagement requirements on fiscal concerns; indeed, the State explained at oral argument that it was not facing such budgetary woes. See Tr. (Complete Transcript Pending). It ultimately does not matter how their argument is characterized; “[it] is both inconsistent with the Medicaid Act and arbitrary and capricious.” Id. at 153. Before briefly restating why that is the case, the Court notes that HHS seems largely to have abandoned any robust form of this argument on appeal. See Stewart v. Azar, Case No. 19-5095, Appellant Brief at 37 (noting only that “it is permissible for HHS to take the optional character of the coverage into account when considering such applications”).

As the Court explained in Stewart II, this position is unpersuasive for three interrelated reasons. To start, Defendants incorrectly assume that “a state has additional discretion to diminish or condition eligibility for the expansion — as opposed to the traditional — population.” 366 F. Supp. 3d at 153. They appear to divine this principle from the Supreme Court’s decision in NFIB, where it found Congress’s decision to require states to expand Medicaid unduly coercive under the Spending Clause. See 567 U.S. at 583–85. But the remedy for this constitutional problem was simply to prevent the Secretary from withdrawing “existing Medicaid funds for failure to comply with the requirements set out in the expansion.” Id. at 585. After the decision, states were thus left with a choice: accept ACA funds and “comply with the conditions on their use” or decline ACA funds and keep prior federal Medicaid appropriations.

22

Id.; see also id. at 586 (explaining that decision does not “affect the Secretary’s ability to withdraw funds provided under the Affordable Care Act if a [s]tate that has chosen to participate in the expansion fails to comply with the requirements of that Act”). Since the Act otherwise places the expansion population on the same plane as the traditional population, states have no “additional discretion” in how to comply with Medicaid requirements as to the expansion group. See Stewart II, 366 F. Supp. 3d at 153.

While Defendants thus err in believing that their treatment of the expansion population is undergirded with any greater discretion than their administration of any other part of the Medicaid program, their argument about flexibility vis-à-vis the expansion population are ultimately a red herring. That is because the entire Medicaid program is optional for states. The Court does not see why — if Defendants are correct that threats to terminate the expansion program can supply the baseline for the Secretary’s § 1115 review — that argument would not be equally good as applied to traditional Medicaid. Id. at 153. Their argument must thus posit that any § 1115 program that maintains any coverage for any set of individuals promotes the objectives of the Medicaid Act as long as the state threatens to terminate all of Medicaid in the absence of waiver approval. The second problem with this position, then, is that it has no limiting principle. Under the Secretary’s reasoning, states may threaten to de-expand, or indeed do away with all of, Medicaid if he does not approve whatever waiver of whatever Medicaid requirements they wish to obtain. The Secretary could then always approve those waivers, no matter how few people remain on Medicaid thereafter, because “any waiver would be coverage promoting compared to a world in which the state offers no coverage at all.” Id. at 154. This reading of the Act would give HHS practically unbridled discretion to implement the Medicaid Act as “an à la carte exercise, picking and choosing which of Congress’s mandates it wishes to

23

implement.” Id. at 153–54. Apart from the potential constitutional concerns such an interpretation would raise, cf. Clinton, 524 U.S. at 440–47 (1998), it clearly constitutes “an impermissible construction of the statute . . . because [it] is utterly unreasonable in [its] breadth.” Aid Ass’n for Lutherans v. U.S. Postal Serv., 321 F.3d 1166, 1178 (D.C. Cir. 2003); see also Agape Church, Inc. v. FCC, 738 F.3d 397, 410 (D.C. Cir. 2013).

The third and final reason to reject this reading is perhaps the most important: it is inconsistent with the text of § 1115. The statute requires the Secretary to evaluate whether the project “is likely to assist in promoting the objectives” of the Act. See 42 U.S.C. § 1315. Against what baseline is he supposed to evaluate the project? The structure of the waiver provision assumes the implementation of the Act. It confirms that the relevant baseline is whether the waiver will promote the objectives of the Act as compared to compliance with the statute’s requirements, “not as compared with a hypothetical future universe” where the Act has no force. Stewart II, 366 F. Supp. 3d at 154. This is so because the overarching provision authorizing these waivers stipulates that, if the Secretary makes a judgment that a demonstration promotes the objectives of the Act, he may then waive compliance with certain of its provisions “to the extent and for the period . . . necessary” to carry out the project. See 42 U.S.C. § 1315(a), (a)(1). That is, the provision contemplates a limited waiver from compliance with the Act’s provisions. Particularly in light of what the Court has discussed above, understanding the baseline as such is the only way this provision makes sense.

For these reasons, the Secretary cannot escape his obligation to consider whether Granite Advantage poses a risk to coverage or is likely to increase coverage by emphasizing the optional nature of the expansion or by citing New Hampshire’s plan to de-expand absent project approval.

24

2. Other Objectives of the Medicaid Act Defendants argue that, regardless of whether the Secretary properly considered Granite Advantage’s effects on coverage, he reasonably approved the project on the ground that it is likely to advance several other Medicaid objectives — namely, the health and financial independence of beneficiaries and the fiscal sustainability of the safety net. HHS relied on those same objectives when it approved Kentucky’s project the second time, and the Court addressed them at length in Stewart II. Indeed, the approval letter reviewed in that decision — which preceded New Hampshire’s by only ten days — tracks the one here practically verbatim with respect to these objectives. Compare AR 1–6 with Kentucky II Approval Letter at 1–6. The agency acknowledged as much during oral argument. See Tr. at 14–15. For the same reasons the Court found that discussion wanting before, it finds it unpersuasive here.

a. Health

The Secretary asserts that “Granite Advantage is [] independently justified because the Secretary found that it was likely to improve the health of Medicaid recipients.” Def. Cross- Motion at 16. Recognizing that this Court has found that health is not a freestanding objective of the Medicaid Act, the agency persists in the contrary view on the ground that “there is little intrinsic value in paying for services if those services are not advancing the health and wellness of individuals receiving them.” Id. This position does not change the Court’s mind. Assuming the Secretary’s interpretation is entitled to Chevron deference, it fails at step two because it falls outside “the bounds of reasonableness.” Abbott Labs v. Young, 920 F.2d 984, 988 (D.C. Cir. 1990). Medicaid, both as enacted and as later expanded by the ACA, reflects Congress’s desire to “mak[e] healthcare more affordable” for “needy populations.” Stewart II, 366 F. Supp. 3d at 144. Congress therefore designed a scheme “to address not health generally but the provision of

25

care to needy populations.” Id. The Secretary cannot “extrapolate the objectives of the statute to a higher level of generality and pursue that aim in the way he prefers.” Id. (citing Waterkeeper Alliance v. EPA, 853 F.3d 527, 535 (D.C. Cir. 2017)).

An example outside of the health-care context helps illustrate the problem with Defendants’ interpretive leap. Say that Congress passed an education statute designed to encourage local control over education because it believed that decentralization was the best way to improve the quality of education. Implementing this view, the statute appropriates money to states that give local school districts control over their own management and curricula. Disagreeing with Congress’s view about local control, however, the Secretary of Education decides she would like to fund experimental state programs that require districts to adhere to strict national education standards. She argues that this advances the purposes of the statute because, in her expertise, national standards improve education, and education must have been the statute’s goal since “there is little intrinsic value in paying” for school expenses “if those services are not advancing” students’ education. See AR 1. This species of argument, of course, would never fly, yet it is indistinguishable from CMS’s in this case, and it exposes the weakness of HHS’s position. By defining a statute’s purposes up a level of generality, the Secretary can justify actions as consistent with the law even if they are clearly at odds with it. This scenario underscores that, as the Court explained in Stewart I, a statute’s objectives are often bound up with the way Congress sought to solve a particular problem. See 313 F. Supp. 3d at 266–67. Ignoring as much grants largely unbounded discretion to agencies, whose exercise of that discretion can veer far afield from anything resembling the statute Congress wrote. Id. at 267– 68.

26

In any event, even assuming Defendants were correct that health is a freestanding objective of the Act, the agency fell short “because it did not consider the health benefits of the project relative to its harms to the health of those who might lose their coverage.” Stewart II, 366 F. Supp. 3d at 145. As with the second Kentucky approval, commenters here made clear that these health effects would be significant. See, e.g., AR 2131–32, 2223–24, 2242–43. The Secretary, moreover, needed to weigh the net effects on health against the effects on coverage more generally. Here he stumbled as well. His neglect to address these considerations is independently fatal to Defendants’ argument that the project is supported by its effects on health alone.

b. Financial Independence Defendants have previously justified the Secretary’s approval on his judgment that community-engagement requirements like those in Granite Advantage improve beneficiaries’ financial independence. See AR 4–5; see also Stewart II, 366 F. Supp. 3d at 145–46. This argument does not appear with much force in the Government’s papers here. To the extent Defendants still press it, the Court rejects it for the reasons discussed at length in Stewart I and Stewart II — namely, that financial independence is not an independent objective of the Act and that the Secretary has not in any event adequately assessed “the benefits of self-sufficiency” and weighed them “against the consequences of coverage loss.” Stewart II, 366 F. Supp. 3d at 148.

c. Fiscal Sustainability

The agency’s principal argument this time around in favor of upholding the Secretary’s decision is the same as it was in Stewart II: that he reasonably concluded that the project would allow “New Hampshire to stretch its limited Medicaid resources.” AR 6. Granite Advantage appears to enhance the fiscal sustainability of the State’s safety net, according to HHS, because

27

beneficiaries who were not previously working may transition to commercial coverage and become healthier. In Stewart II, the Court agreed that fiscal sustainability was a valid consideration in a Section 1115 project, but it found the Secretary’s explanation for why the project addressed that concern to be arbitrary and capricious. See 366 F. Supp. 3d at 149–52. As the agency recognized at oral argument, HHS’s explanation in approving Granite Advantage, which is practically identical to what it said in Stewart II, likewise cannot clear the bar. See Tr. at 14–15.

To start, HHS made no finding that Granite Advantage would save the state “any amount of money or otherwise make the program more sustainable in some way.” Id. at 149. With respect to savings, the Court is not suggesting that the Secretary “must quantify some exact amount . . . , but he must make some finding that supports his conclusion that the project” addresses New Hampshire’s fiscal concerns. Id. at 149–50. Two considerations make this analysis especially important in this case.

First, the State has represented that it neither intends for the demonstration to reduce costs nor expects it to do so. At oral argument, it explained that New Hampshire is not encountering the same fiscal concerns as Kentucky with respect to its Medicaid program and that reducing health costs is not in fact an objective of this demonstration project. See Tr. (Full Transcript Pending). Consistent with that position, New Hampshire’s waiver application projected “that spending growth in the future [under Granite Advantage] will be consistent with standard growth rates experienced in the past.” AR 4399. The glaring disconnect between the Secretary’s position and New Hampshire’s raises substantial questions about how the agency came to believe the program would improve the State’s fiscal circumstances, underscoring the need for reasoned analysis of this issue.

28

Second, the record in this case contains substantial reasons to doubt whether the program will save any money given administrative costs and the possible rise in uncompensated care that would accrue to the State. See AR 1480, 1949, 2206, 2241–43, 2534, 2710–11; cf. 42 U.S.C. § 1396d(y)(1) (providing that federal government shall pay between 90 and 100 percent of costs to expansion population). While the Secretary may well have arrived at a different conclusion, he needed to explain how he got there in light of the nearly uniform evidence going the other direction.

The agency does propose several potential mechanisms by which the program could save the State money. But those do not advance the ball because they are conclusory and unsupported by any evidence in the record. The Secretary never explained, for example, why he thinks that the program will transition beneficiaries to commercial coverage, given the consistent evidence before him that nearly all Medicaid recipients are already working, unable to work, or able to find only low-paying jobs that do not offer or lead to commercial coverage. See AR 1949, 2209–10, 2225, 2435; see also id. at 3587 (explaining that work requirements only expected to apply to 6% of able and non-working beneficiaries and that work opportunities for them are limited). That the community-engagement requirements can be met through education or volunteer activities, rather than employment, gives more reason to wonder why the Secretary thought the program would expand access to commercial coverage. Likewise, the agency has not explained the sort of health benefits it expected would accrue to beneficiaries as a result of the new requirements and how such benefits would save the State money.

Apart from these failings, the Secretary’s fiscal-sustainability discussion suffers from another key flaw: it did not “compare the benefit[s]” to the State’s safety net “to the consequences for coverage.” Stewart II, 366 F. Supp. 3d at 150. The D.C. Circuit’s decision in

29

Pharmaceutical Research & Manufacturers of America v. Thompson, 362 F.3d 817 (D.C. Cir. 2004), helps illustrate how the agency fell short in this respect. There, the court upheld HHS’s decision to impose a minor restriction on beneficiaries’ access to prescription drugs to try to prevent borderline populations from becoming Medicaid eligible. Id. at 825. In doing so, the D.C. Circuit discussed at length whether the agency had adequately considered the burden on Medicaid recipients and reasonably explained why such imposition was necessary under the circumstances. Id. at 825–26. This is just the kind of analysis the Secretary never conducted here. He neither addressed the magnitude of coverage loss nor weighed that concern against the asserted fiscal-sustainability benefits.

This analysis is essential given the admonitions in Thompson and the Supreme Court’s decision in Pharmaceutical Research & Manufacturers of America v. Walsh, 538 U.S. 644 (2003) (plurality), that projects imposing significant burdens on Medicaid recipients may not be consistent with the Act’s purposes. Id. at 664–65 (“The fact that the [Program] . . . provid[es] benefits to needy persons and . . . curtail[s] the State’s Medicaid costs . . . would not provide a sufficient basis for upholding the program if it severely curtailed Medicaid recipients’ access to prescription drugs.”); Thompson, 362 F.3d at 826 (relying on “the absence of any demonstrable significant impediment to Medicaid services”). As explained in Stewart II, “That there are limits on the extent to which fiscal sustainability can justify cuts like those outlined in these cases makes sense. Most cuts to Medicaid services would reduce the cost of Medicaid and thus advance the sustainability of the program to some extent. But it would be nonsensical to conclude that any cut therefore always promotes the Act’s objectives.” 366 F. Supp. 3d at 152.

In short, the Court finds that the Secretary’s assertion that Granite Advantage advances the purposes of Medicaid because it would improve the fiscal sustainability of New Hampshire’s

30

Medicaid program is arbitrary and capricious, especially where the State at argument generally disclaimed such motivation. In so concluding, the Court is not questioning the agency’s predictive judgments or evaluating the evidence before it on this issue. It is simply looking for what the APA requires: a reasoned explanation that considers the factors relevant to the agency’s decision.

d. Data Collection

The Secretary has one more arrow in his quiver. He suggests that Granite Advantage advances the objectives of Medicaid regardless of what effect it has on beneficiary health and coverage because it helps the State and agency collect useful data for future policymaking purposes. See Def. Cross-Mot. at 22–3; AR 12. This one holds no water. As a textual matter, the Secretary is authorized to approve only those projects “likely to assist in promoting the objectives” of the Medicaid Act. See 42 U.S.C. § 1315(a). But no one is suggesting with a straight face that a purpose of the Medicaid Act is to collect data. A demonstration can hardly be justified, therefore, solely on data-collection grounds. The practical consequences of the Government’s suggestion are also alarming. If experimentation alone could justify a project, then demonstrations with dire consequences for Medicaid beneficiaries could be approved just for the Government to gather information. Recognizing these difficulties, HHS disclaimed any such position at oral argument, confirming that any demonstration project must do more than collect data: it must also advance the purposes of the Medicaid Act, including the core objective of providing medical assistance to the needy. See Tr. at 14.

31

Remedy

That leaves only consideration of the proper remedy. Three sets of issues require attention.

First, the Court must decide whether, having arrived at the conclusions outlined above, it should issue its decision now or instead wait until November, nearer in time to when the community-engagement requirements are set to take effect. HHS, at oral argument, maintained that the Court should stay its pen because vacating the program will substantially disrupt New Hampshire’s implementation and outreach efforts and because the Court of Appeals may issue a decision in the Kentucky and Arkansas cases in the interim. See Tr. at 4–6. Neither factor, however, counsels in favor of delay. As to the former, the State of New Hampshire, which is presumably best situated to understand the consequences of a timely decision, asks the Court to issue its Opinion now rather than waiting until November. It explained at oral argument that a decision would provide further clarity to the State while it compiles an amended waiver application and considers how to implement its program moving forward. Id. (Complete Transcript Pending). Citing the uncertainty attending their circumstances and those of other Medicaid beneficiaries, Plaintiffs agree that the Court should not wait to act. As to the latter, while it is possible that the Court of Appeals will issue its decision between now and November, it is also possible that the Circuit’s decision will post-date the December 1 implementation. At this point, no oral-argument date has been set. Either way, that consideration does not outweigh New Hampshire’s and Plaintiffs’ interests in seeking a timelier decision.

Second, the Court must consider whether to remand this matter to the agency without vacating its underlying approval of New Hampshire’s demonstration. When a court concludes that agency action is unlawful, “the practice of the court is ordinarily to vacate the rule.” Ill.

32

Pub. Telecomms. Ass’n v. FCC, 123 F.3d 693, 693 (D.C. Cir. 1997); Sierra Club v. Van Antwerp, 719 F. Supp. 2d 77, 78 (D.D.C. 2010) (“[B]oth the Supreme Court and the D.C. Circuit Court have held that remand, along with vacatur, is the presumptively appropriate remedy for a violation of the APA.”). Remand without vacatur may be appropriate, however, depending on the “seriousness of the order’s deficiencies (and thus the extent of doubt whether the agency chose correctly) and the disruptive consequences of an interim change.” Allied-Signal, Inc. v. U.S. Nuclear Regulatory Comm’n, 988 F.2d 146, 150–51 (D.C. Cir. 1993) (citation omitted); see also Standing Rock Sioux Tribe v. U.S. Army Corps of Engineers, 282 F. Supp. 3d 91, 103 (D.D.C. 2017) (declining to vacate when agency “largely complied” with statute and could likely substantiate prior conclusions on remand).

In Stewart I, Stewart II, and Gresham, the Court concluded that both factors supported vacatur. The Secretary’s failure to consider an objective of Medicaid is a “major shortcoming” going “to the heart” of his decisions. See 313 F. Supp. 3d at 273. As to Kentucky HEALTH, vacatur was not overly disruptive because the project had “yet to take effect” and the plaintiffs could suffer “serious harm[s]” were Kentucky HEALTH allowed to be implemented pending further proceedings. Id.; see also Stewart II, 366 F. Supp. 3d at 156. And as to Arkansas, while vacatur would concededly have been disruptive given that the program had already begun in part, the seriousness of this disruption — which was largely administrative in nature — had to be “balanced against the harms that Plaintiffs and persons like them will experience if the program remains in effect.” Gresham, 363 F. Supp. 3d at 184.

Defendants do not dispute that vacatur is appropriate under this analysis. See HHS Cross-Mot. at 27–29 (declining to argue that Allied-Signal factors counsel against vacatur). And for good reason. Like the prior three approvals, this one suffers from a significant deficiency:

33

the failure to address a central factor in its decision. And vacatur will not be terribly disruptive, given that New Hampshire has not fully implemented the community-engagement requirements — indeed, the State has continued to make legislative and executive tweaks to the program since CMS approved it in late 2018.

Third, HHS argues that rather than vacating the November 2018 approval as a whole, this Court should tailor any relief solely to the four Plaintiffs and the aspects of the program that they have successfully challenged. Id. at 27–29. The Court is no more persuaded by this contention than it was in Stewart II. See 366 F. Supp. 3d at 155. In an APA case, the “ordinary result” of the Court’s finding an agency action unlawful is to vacate that action — not to judicially re-write what the agency did so that it somehow does not apply to a narrow group of people or so that it persists piecemeal. See Harmon v. Thornburgh, 878 F.2d 484, 495 n.21 (D.C. Cir. 1989). The Court of Appeals, accordingly, has explained that “if the plaintiff prevails” in challenging an agency action under the APA, “the result is that the rule is invalidated, not simply that the court forbids its application to a particular individual.” Nat’l Min. Ass’n v. U.S. Army Corps of Engineers, 145 F.3d 1399, 1409 (D.C. Cir. 1998) (quoting Lujan v. Nat’l Wildlife Fed., 497 U.S. 871, 913 (1990) (Blackmun, J., dissenting)).

Defendants offer no reason to deviate from the settled manner in which courts accord relief in APA cases; indeed, they do not cite a single APA case in which relief has been granted in this manner. Plaintiffs have standing to challenge and have successfully challenged the Secretary’s November 30, 2018, approval of the amendments to Granite Advantage. The proper relief, under these circumstances, is vacatur of that action. IV. CONCLUSION For the foregoing reasons, the Court will grant Plaintiffs’ Motion for Summary Judgment

34

and deny Defendants’ Cross-Motion and Motion to Dismiss. A separate Order consistent with this Opinion will issue this day.

/s/ James E. Boasberg

JAMES E. BOASBERG

United States District Judge

Date: July 29, 2019

APPENDIX A

Comments on Coverage Consequences of Community-Engagement Requirements AR 1484 (Disability Rights Center) (“Red tape and paperwork requirements have been shown to reduce enrollment in Medicaid and individuals coping with serious mental illness or physical impairments may face particular difficulty meeting these requirements.”); AR 1489 (American Cancer Society Cancer Action Network) (“If work and community engagement is required as a condition of eligibility, many cancer patients, recent survivors, and those with other chronic illnesses could find that they are ineligible for the lifesaving care and treatment services provided through the state’s Medicaid program . . . The increase in administrative requirements for enrollees to attest to their working status would likely decrease the number of individuals with Medicaid coverage, regardless of whether they are exempt.”); AR 1949 (Rights and Democracy NH) (“[I]mposing work requirements for people who need Medicaid Expansion to gain access to health care will likely cause many people to lose coverage and increase levels of uncompensated care.”); AR 1953 (Cystic Fibrosis Foundation) (“[M]any people may have trouble complying with new eligibility requirements and for someone with [cystic fibrosis], this could result in a life-threatening gap in coverage.”); AR 1956 (NH Community Behavioral Health Association) (“We fear that the work requirement has the very real potential to jeopardize care for individuals with mental illness if they lose their Medicaid coverage. . . . Asking people struggling with mental illness to document their work by keeping track of every week’s pay stubs is an onerous requirement.”); AR 2132 (American Diabetes Association) (“New Hampshire’s proposal to limit access to Medicaid services through the implementation of work requirements will decrease access to care for low-income New Hampshire residents with and at risk for diabetes and increase state health care costs.”); AR 2201 (Lung Cancer Alliance) (citing estimates that similar Iowa waiver caused that “more than 3,000 beneficiaries” to lose coverage and “become uninsured”); AR 2208 (Center on Budget and Policy Priorities & Georgetown University Center for Children and Families) (“Kaiser conservatively estimates that between 25 and 50 percent of such enrollees are at risk of losing coverage under work requirements. New Hampshire’s work requirement is especially onerous, because just one month of non-compliance can lead to loss of coverage.”); AR 2224–25 (American Heart Association) (asserting that work requirements will reduce “access to healthcare services both in the short and long term” for people with cardiovascular disease); AR 2238 (National Council on Aging) (“Extending work requirements would also particularly hurt the rural residents of NH,” who make up 48% of NH’s Medicaid population, because they

35

are less likely to have access to the internet or transportation and therefore “risk losing coverage.”); AR 2241 (Center for Law and Social Policy) (“The evidence from other programs with similar requirements is that in spite of official exemptions, in practice, individuals with disabilities are often not exempted and are more likely to lose benefits.”); AR 2434 (CLASP Volatile Job Schedules and Access to Public Benefits Report) (finding that tying benefits to work requirements, especially a specific number of hours, can be problematic because “more than 40 percent of early career hourly-workers (ages 26 to 32) receive one week or less advance notice of their job schedules. Half of these workers have no input into their schedules and three-quarters experience fluctuations in the number of hours they work, with hours varying by more than eight hours per week on average.”); AR 2530 (Kaiser Family Foundation Issue Brief) (considering national effects of a Medicaid work requirement and concluding that, “[o]verall, among the 23.5 million non-SSI, non-dual, nonelderly Medicaid adults, disenrollment ranges from 1.4 million to 4.0 million under the scenarios considered”); AR 2563 (New Futures) (stating that results of Arkansas’s similar program will likely reflect those of NH’s program, and “[i]n the first month of implementation of the Arkansas Works work requirement (June 2018) fewer than six percent of the nearly 8,000 Medicaid enrollees who did not declare an exemption were able to satisfy the reporting requirement”); AR 2575 (Families USA) (“The coverage losses [caused by the work requirement] will result in an increase in the state’s uninsured population, lost health care access, and worse health for low- income adults in New Hampshire.”); AR 2696 (National Health Law Program) (“All evidence indicates that the work requirement will lead to substantial numbers of individuals losing Medicaid coverage. In Kentucky, which proposed a similar work requirement, researchers have estimated that between 45,000 to 103,000 individuals could lose coverage due to the work requirement alone.”); AR 2963 (New Hampshire Fiscal Policy Institute) (“Work requirements implemented in Arkansas for certain populations starting in June 2018 . . . provide an initial indication that similar requirements may reduce enrollment significantly in New Hampshire.”); AR 3372 (Kaiser Family Foundation) (“Under the Medicaid work requirement programs, the population subject to Medicaid work requirements may have access to only low-wage, unstable, or low-quality jobs to meet the weekly/monthly hours requirement[;] . . . ‘[p]olicies that promote job growth without giving attention to the overall adequacy of the jobs may undermine health.’”); AR 3406 (Urban Institute) (“The red tape associated with work requirements can cause people to lose access to vital supports even when they are working or should be exempt from the requirements.”); AR 3584 (Philip Rocco, PhD) (citing a study about “a Florida welfare reform experiment whose benefits were conditioned on workforce participation [and] had a 16 percent higher mortality rate than comparable recipients of welfare who were not subject to work stipulations”); AR 3644 (New Hampshire Legal Assistance) (“All evidence indicates that the work requirement will lead to substantial numbers of individuals losing Medicaid coverage.”); AR 4563–64 (Leukemia & Lymphoma Society) (“New Hampshire’s proposal to extend its work requirement will perpetuate a return to increased bureaucracy and paperwork and, in turn, coverage losses.”).